Review normalize command architecture

.gitignore

@@ -1,3 +1,7 @@
+# ---> Codex
+.codex
+AGENTS.md
+
 # ---> Go
 # If you prefer the allow list template instead of the deny list, see community template:
 # https://github.com/github/gitignore/blob/main/community/Golang/Go.AllowList.gitignore

.woodpecker/release.yml

@@ -8,7 +8,7 @@ steps:
       - |
         set -eu
 
-        version="${CI_COMMIT_TAG}"
+        version="$CI_COMMIT_TAG"
         dist="dist"
         pkg="gitea.maximumdirect.net/eric/seriatim/cmd/seriatim"
 
@@ -19,10 +19,10 @@ steps:
           goos="$1"
           goarch="$2"
           suffix="$3"
-          output="$dist/seriatim_${version}_${goos}_${goarch}${suffix}"
+          output="$dist/seriatim-$version-$goos-$goarch$suffix"
 
           CGO_ENABLED=0 GOOS="$goos" GOARCH="$goarch" \
-            go build -trimpath -ldflags "-s -w -X gitea.maximumdirect.net/eric/seriatim/internal/buildinfo.Version=${version}" \
+            go build -trimpath -ldflags "-s -w -X gitea.maximumdirect.net/eric/seriatim/internal/buildinfo.Version=$version" \
             -o "$output" "$pkg"
         }
 
@@ -33,11 +33,6 @@ steps:
         build_binary windows amd64 ".exe"
         build_binary windows arm64 ".exe"
 
-        (
-          cd "$dist"
-          sha256sum seriatim_* > SHA256SUMS
-        )
-
   - name: publish-release
     image: woodpeckerci/plugin-release
     depends_on:
@@ -46,8 +41,10 @@ steps:
       api_key:
         from_secret: GITEA_RELEASE_TOKEN
       files:
-        - dist/seriatim_*
-        - dist/SHA256SUMS
+        - dist/seriatim-*
+      checksum: sha256
+      checksum-file: SHA256SUMS
+      checksum-flatten: true
       file-exists: skip
       overwrite: false
       prerelease: false

README.md

@@ -1,8 +1,8 @@
 # seriatim
 
-`seriatim` merges per-speaker WhisperX-style JSON transcripts into a single JSON transcript that preserves speaker identity and chronological order.
+`seriatim` merges per-speaker WhisperX-style JSON transcripts into a single JSON transcript that preserves speaker identity and chronological order. It also trims existing seriatim output artifacts by segment ID and normalizes external transcript-like JSON into standard seriatim output schemas.
 
-The current implementation supports the `merge` command. It reads one or more input JSON files, optionally maps each input file to a canonical speaker using `speakers.yml`, sorts all segments by timestamp, detects and resolves overlaps when word-level timing is available, assigns consecutive numeric `id` values, and writes a merged JSON artifact.
+The current implementation supports the `merge`, `trim`, and `normalize` commands. `merge` reads one or more input JSON files, optionally maps each input file to a canonical speaker using `speakers.yml`, sorts all segments by timestamp, detects and resolves overlaps when word-level timing is available, assigns consecutive numeric `id` values, and writes a merged JSON artifact. `trim` reads an existing seriatim output artifact and projects it to a retained segment subset. `normalize` reads transcript-like JSON input, validates required segment fields, sorts deterministically, assigns fresh IDs, and emits a selected seriatim output schema.
 
 ## Usage
 
@@ -25,10 +25,39 @@ go run ./cmd/seriatim merge \
   --report-file report.json
 ```
 
+Trim an existing seriatim artifact:
+
+```sh
+go run ./cmd/seriatim trim \
+  --input-file merged.json \
+  --output-file trimmed.json \
+  --keep "1-10, 15, 20-25"
+```
+
+Normalize external transcript-style JSON:
+
+```sh
+go run ./cmd/seriatim normalize \
+  --input-file transcript.json \
+  --output-file normalized.json
+```
+
+Normalize an Audita-style bare segment array to full schema with report output:
+
+```sh
+go run ./cmd/seriatim normalize \
+  --input-file audita-segments.json \
+  --output-file normalized-full.json \
+  --output-schema seriatim-full \
+  --report-file normalize-report.json
+```
+
 ## CLI
 
 ```text
 seriatim merge [flags]
+seriatim trim [flags]
+seriatim normalize [flags]
 ```
 
 Global flags:
@@ -49,16 +78,98 @@ Global flags:
 | `--autocorrect` | No | none | Autocorrect rules YAML file. When omitted, the default `autocorrect` module leaves text unchanged. |
 | `--input-reader` | No | `json-files` | Input reader module. |
 | `--output-modules` | No | `json` | Comma-separated output modules. |
-| `--output-schema` | No | `default` | JSON output contract. Allowed values are `default`, `minimal`, and `seriatim`. |
+| `--output-schema` | No | `seriatim-intermediate` | JSON output contract. Allowed values are `seriatim-minimal`, `seriatim-intermediate`, and `seriatim-full`. If omitted, the runtime default is used; consumers that depend on a specific shape should set this explicitly. |
 | `--preprocessing-modules` | No | `validate-raw,normalize-speakers,trim-text` | Comma-separated preprocessing modules, evaluated in order. |
 | `--postprocessing-modules` | No | `detect-overlaps,resolve-overlaps,backchannel,filler,resolve-danglers,coalesce,detect-overlaps,autocorrect,assign-ids,validate-output` | Comma-separated postprocessing modules, evaluated in order. |
 | `--coalesce-gap` | No | `3.0` | Maximum same-speaker gap in seconds for `coalesce`; also used as the `resolve-overlaps` context window. Must be a non-negative float. |
 
+`trim` flags:
+
+| Flag | Required | Default | Description |
+| --- | --- | --- | --- |
+| `--input-file` | Yes | none | Input seriatim output artifact JSON file. |
+| `--output-file` | Yes | none | Trimmed transcript JSON output path. |
+| `--keep` | Exactly one of `--keep` or `--remove` is required | none | Segment ID selector to retain. |
+| `--remove` | Exactly one of `--keep` or `--remove` is required | none | Segment ID selector to drop. |
+| `--output-schema` | No | preserve input artifact schema | Optional output schema override: `seriatim-minimal`, `seriatim-intermediate`, or `seriatim-full`. |
+| `--report-file` | No | none | Optional report JSON output path. |
+| `--allow-empty` | No | `false` | Allow trimming to zero retained segments. |
+
+`trim` selection rules:
+
+- `--keep` and `--remove` are mutually exclusive.
+- Exactly one of `--keep` or `--remove` is required.
+- Selection is by segment ID only.
+- Invalid selected segment IDs fail the command by default.
+
+`trim` selector syntax:
+
+- Segment IDs are positive 1-based integers.
+- Inclusive ranges are supported: `1-10`.
+- Comma-separated selectors are supported: `1-10,15,20-25`.
+- Whitespace around numbers, commas, and hyphens is allowed: `1 - 10, 15, 20 - 25`.
+- Duplicate and overlapping ranges are accepted and normalized as a union.
+- Descending ranges (for example `10-1`) are rejected.
+
+`trim` behavior:
+
+- `trim` consumes existing seriatim JSON output artifacts only.
+- `trim` does not accept raw WhisperX transcript JSON as input.
+- Retained output segment IDs are renumbered sequentially from `1` to `N`.
+- Transcript order is preserved from input transcript order; selector order does not reorder output.
+- When output schema is `seriatim-full`, overlap groups are recomputed from retained segments.
+- `--output-schema seriatim-full` is supported when trim has full-schema artifact data to emit; trim does not synthesize missing full-schema provenance from minimal/intermediate input artifacts.
+- `trim` does not run merge postprocessors such as `resolve-overlaps`, `coalesce`, or `autocorrect`.
+
+`trim` report output:
+
+- When `--report-file` is provided, the report includes standard trim/validation/output events.
+- The report includes a `trim-audit` event containing trim operation metadata, including selected IDs, retained/removed counts, removed IDs, and old-to-new segment ID mapping.
+- Old-to-new ID mapping is emitted as a deterministic ordered array of `{old_id, new_id}` pairs.
+
+`normalize` flags:
+
+| Flag | Required | Default | Description |
+| --- | --- | --- | --- |
+| `--input-file` | Yes | none | Input transcript JSON file. |
+| `--output-file` | Yes | none | Normalized transcript JSON output path. |
+| `--output-schema` | No | `seriatim-intermediate` (resolved via `SERIATIM_OUTPUT_SCHEMA` when set) | Output JSON schema: `seriatim-minimal`, `seriatim-intermediate`, or `seriatim-full`. |
+| `--output-modules` | No | `json` | Comma-separated output modules. Current normalize support is `json` only. |
+| `--report-file` | No | none | Optional report JSON output path. |
+
+`normalize` input shapes:
+
+- Top-level object with a `segments` array.
+- Bare top-level array of segment objects (for example, Audita-style output).
+
+`normalize` required segment fields:
+
+- `start`
+- `end`
+- `speaker`
+- `text`
+
+`normalize` behavior:
+
+- Validates `start >= 0`, `end >= start`, and non-empty `speaker`.
+- Accepts existing input `id` values as provenance only.
+- Reassigns output segment IDs sequentially from `1` to `N`.
+- Sorts deterministically by `(start, end, original_input_index, speaker)`.
+- Uses original input order only as a tie-breaker.
+- Does not run merge postprocessors such as overlap detection, overlap resolution, coalescing, or autocorrect.
+- Useful for converting external transcript outputs into standard seriatim artifacts.
+
+`normalize` report output:
+
+- When `--report-file` is provided, normalize emits deterministic report events with input shape detection, segment counts, schema/module selections, sorting/ID diagnostics, and output write/validation summaries.
+- A machine-readable `normalize-audit` event is included for downstream tooling.
+
 Environment variables:
 
 | Environment Variable | Default | Description |
 | --- | --- | --- |
-| `SERIATIM_OVERLAP_WORD_RUN_GAP` | `0.75` | Maximum gap in seconds between adjacent timed words when `resolve-overlaps` builds word-run replacement segments. Must be a positive float. |
+| `SERIATIM_OUTPUT_SCHEMA` | `seriatim-intermediate` | Output schema used when `--output-schema` is not explicitly provided. Allowed values are `seriatim-minimal`, `seriatim-intermediate`, and `seriatim-full`. The CLI flag takes precedence. |
+| `SERIATIM_OVERLAP_WORD_RUN_GAP` | `1.0` | Maximum gap in seconds between adjacent timed words when `resolve-overlaps` builds word-run replacement segments. Must be a positive float. |
 | `SERIATIM_OVERLAP_WORD_RUN_REORDER_WINDOW` | `1.0` | Near-start window in seconds for ordering replacement word runs shortest-first. Must be a positive float. |
 | `SERIATIM_BACKCHANNEL_MAX_DURATION` | `2.0` | Maximum duration in seconds for `backchannel` classification. Must be a positive float. |
 | `SERIATIM_FILLER_MAX_DURATION` | `1.25` | Maximum duration in seconds for `filler` classification. Must be a positive float. |
@@ -159,14 +270,16 @@ The old `inputs:` direct mapping format is no longer supported.
 
 `--output-modules json` controls the writer. `--output-schema` controls the JSON contract that writer serializes.
 
-The `default` schema is the default output contract. It stays close to `minimal`, but adds optional `categories` on each segment:
+The named schemas are stable public contracts. If a consumer depends on a specific shape, it should request that schema explicitly at runtime. The runtime default selection may change in a future release.
+
+The `seriatim-intermediate` schema is the current default selection when neither `--output-schema` nor `SERIATIM_OUTPUT_SCHEMA` is set. It stays close to the minimal schema, but adds optional `categories` on each segment:
 
 ```json
 {
   "metadata": {
     "application": "seriatim",
     "version": "dev",
-    "output_schema": "default"
+    "output_schema": "seriatim-intermediate"
   },
   "segments": [
     {
@@ -181,7 +294,7 @@ The `default` schema is the default output contract. It stays close to `minimal`
 }
 ```
 
-The explicit `seriatim` schema uses the full seriatim envelope:
+The `seriatim-full` schema uses the full seriatim envelope:
 
 ```json
 {
@@ -191,7 +304,7 @@ The explicit `seriatim` schema uses the full seriatim envelope:
     "input_reader": "json-files",
     "input_files": ["eric.json", "mike.json"],
     "preprocessing_modules": ["validate-raw", "normalize-speakers", "trim-text"],
-    "postprocessing_modules": ["detect-overlaps", "resolve-overlaps", "backchannel", "filler", "coalesce", "resolve-danglers", "detect-overlaps", "autocorrect", "assign-ids", "validate-output"],
+    "postprocessing_modules": ["detect-overlaps", "resolve-overlaps", "backchannel", "filler", "resolve-danglers", "coalesce", "detect-overlaps", "autocorrect", "assign-ids", "validate-output"],
     "output_modules": ["json"]
   },
   "segments": [
@@ -231,14 +344,14 @@ The explicit `seriatim` schema uses the full seriatim envelope:
 }
 ```
 
-The `minimal` schema emits minimal metadata and compact ordered segments:
+The `seriatim-minimal` schema emits minimal metadata and compact ordered segments:
 
 ```json
 {
   "metadata": {
     "application": "seriatim",
     "version": "dev",
-    "output_schema": "minimal"
+    "output_schema": "seriatim-minimal"
   },
   "segments": [
     {
@@ -254,7 +367,7 @@ The `minimal` schema emits minimal metadata and compact ordered segments:
 
 Minimal output intentionally omits categories, overlap groups, source/provenance fields, and pipeline configuration metadata.
 
-Default output intentionally omits overlap groups and source/provenance fields, but keeps optional `categories` and minimal metadata.
+Intermediate output intentionally omits overlap groups and source/provenance fields, but keeps optional `categories` and minimal metadata.
 
 Segments are sorted deterministically by:
 
@@ -270,7 +383,7 @@ The public Go output contract is available from:
 import "gitea.maximumdirect.net/eric/seriatim/schema"
 ```
 
-The same package embeds machine-readable JSON Schemas in `schema/output.schema.json`, `schema/default-output.schema.json`, and `schema/minimal-output.schema.json`. The default `validate-output` postprocessor validates the selected output shape and verifies final segment IDs are present, sequential, and start at `1`.
+The same package embeds machine-readable JSON Schemas in `schema/full-output.schema.json`, `schema/intermediate-output.schema.json`, and `schema/minimal-output.schema.json`. The default `validate-output` postprocessor validates the selected output shape and verifies final segment IDs are present, sequential, and start at `1`.
 
 ## Overlap Detection
 
@@ -295,12 +408,12 @@ For each detected overlap group, `resolve-overlaps` uses preserved WhisperX word
 
 - The resolution window expands the detected overlap group by `--coalesce-gap` seconds on both sides.
 - Nearby same-speaker context segments are included when they intersect the expanded window and their start or end is within `--coalesce-gap` of the original overlap boundary.
-- Words are included when their interval intersects the expanded resolution window.
+- Once a segment is selected for replacement, all timed words from that segment participate in word-run construction; the window controls segment selection, not per-word clipping.
 - Context segments that are part of another detected overlap group are not pulled into the current group.
 - Untimed words are included in replacement text in original word order when nearby timed words create a replacement run.
 - Untimed words do not affect replacement segment start/end times or word-run gap splitting.
 - Words for the same speaker are merged into one run when the gap between adjacent words is no greater than `SERIATIM_OVERLAP_WORD_RUN_GAP`.
-- The default word-run gap is `0.75` seconds.
+- The default word-run gap is `1.0` seconds.
 - Set `SERIATIM_OVERLAP_WORD_RUN_GAP` to a positive number of seconds to override the default.
 - Near-start replacement word runs are reordered so shorter segments come first when adjacent starts are within `SERIATIM_OVERLAP_WORD_RUN_REORDER_WINDOW`.
 - The default word-run reorder window is `1.0` seconds.
@@ -339,12 +452,12 @@ The default pipeline runs `resolve-danglers` before `coalesce` and before the se
 
 - Dangling-end fragments have no more than two words and end in punctuation.
 - Dangling-start fragments have no more than two words.
-- Matching uses any shared `derived_from` value.
+- Matching uses same-speaker segments with any shared `derived_from` value.
 - Merged segments use `source_ref` values such as `resolve-danglers:1`, keep the target segment's transcript position, and union `derived_from`.
 
 ## Coalescing
 
-The default pipeline runs `coalesce` after `resolve-danglers` and the second overlap detection pass. It merges adjacent same-speaker segments in the transcript's current order when `next.start - current.end <= --coalesce-gap`.
+The default pipeline runs `coalesce` after `resolve-danglers` and before the second overlap detection pass. It merges adjacent same-speaker segments in the transcript's current order when `next.start - current.end <= --coalesce-gap`.
 
 Coalesced segments use `source_ref` values such as `coalesce:1`, include `derived_from`, and omit `source_segment_index`.
 

architecture.md

@@ -1,6 +1,10 @@
 # seriatim Architecture
 
-`seriatim` is a deterministic transcript merge utility for combining multiple per-speaker transcript inputs into a single chronologically ordered diarized transcript.
+`seriatim` is a deterministic transcript utility for:
+
+- merging multiple per-speaker transcript inputs into a single chronologically ordered diarized transcript, and
+- projecting existing seriatim transcript artifacts through deterministic segment-ID trimming, and
+- canonicalizing external transcript-style JSON inputs into standard seriatim output schemas.
 
 The initial use case is merging independently transcribed speaker audio tracks from the same recorded session, such as a weekly tabletop RPG session. The architecture should also support meetings, podcasts, interviews, and other multi-speaker events.
 
@@ -20,6 +24,7 @@ The initial use case is merging independently transcribed speaker audio tracks f
 8. Detect and annotate overlapping speech regions.
 9. Emit one or more output artifacts through output writers.
 10. Produce report data for validation findings, corrections, and transformations.
+11. Support artifact-level transcript projection commands that operate on existing seriatim output.
 
 ## Non-goals
 
@@ -56,6 +61,8 @@ configuration check
 
 Each stage has an explicit data contract. Input and output stages perform I/O. Processing stages should be deterministic transformations over in-memory models and should record report events for validation findings, corrections, and transformations.
 
+`merge` runs this pipeline. `trim` and `normalize` are intentionally separate from this pipeline and operate at the artifact layer.
+
 ## Stage Contracts
 
 ### 1. Configuration Check
@@ -173,6 +180,14 @@ The current output format is JSON, specified with:
 --output-file merged.json
 ```
 
+The current named JSON schemas are:
+
+- `seriatim-minimal`
+- `seriatim-intermediate`
+- `seriatim-full`
+
+The current runtime default selection is `seriatim-intermediate`, but default selection may change over time. Consumers that depend on a specific schema should request it explicitly.
+
 Future output formats may include:
 
 - Markdown.
@@ -183,6 +198,41 @@ Future output formats may include:
 
 Output writers should be selected from an explicit registry and should consume the final transcript model read-only. Multiple output writers may run for a single invocation.
 
+### 7. Artifact Projection Stage (`trim` command)
+
+`trim` is an artifact-level command that reads an existing seriatim output artifact and emits a projected artifact containing a segment-ID subset.
+
+Design constraints:
+
+- `trim` runs after `merge`, not as a merge postprocessor.
+- `trim` validates the input artifact against supported seriatim output schemas.
+- `trim` performs deterministic keep/remove selection by segment ID.
+- `trim` renumbers retained IDs to `1..N` in transcript order.
+- `trim` validates the final output against the selected output schema before writing.
+- `trim` records audit metadata in report output.
+
+`trim` is intentionally separate from merge postprocessing because it consumes already-emitted public artifacts. This separation keeps merge semantics stable and avoids rerunning merge-only transforms on projected artifacts.
+
+`trim` must not rerun merge postprocessors such as `resolve-overlaps`, `coalesce`, or `autocorrect`.
+
+### 8. Artifact Canonicalization Stage (`normalize` command)
+
+`normalize` is an artifact-level command that reads transcript-like JSON and emits a standard seriatim output artifact in a selected schema.
+
+Design constraints:
+
+- `normalize` runs outside the merge pipeline and does not invoke merge preprocessing or postprocessing modules.
+- `normalize` accepts two input shapes: object-with-`segments` and bare segment arrays.
+- `normalize` validates required segment fields (`start`, `end`, `speaker`, `text`) and timing/speaker constraints.
+- `normalize` sorts segments deterministically by chronological keys and stable input-index tie-breakers.
+- `normalize` assigns fresh sequential output IDs (`1..N`) after sorting.
+- `normalize` validates final output against the selected schema before writing.
+- `normalize` writes optional deterministic report diagnostics when `--report-file` is requested.
+
+`normalize` is intended for canonicalizing external transcript outputs (including Audita-style bare arrays) into seriatim contracts, not for running merge-time language or overlap transformations.
+
+`normalize` must not run merge postprocessors such as overlap detection, overlap resolution, coalescing, or autocorrect.
+
 ## Module Classification
 
 Modules should be classified by their contract and allowed effects.
@@ -216,7 +266,7 @@ seriatim merge \
   --preprocessing-modules validate-raw,normalize-speakers,trim-text \
   --postprocessing-modules detect-overlaps,resolve-overlaps,backchannel,filler,resolve-danglers,coalesce,detect-overlaps,autocorrect,assign-ids,validate-output \
   --output-modules json \
-  --output-schema default \
+  --output-schema seriatim-intermediate \
   --output-file merged.json \
   --report-file report.json
 ```
@@ -358,7 +408,7 @@ Initial classifications may include:
 - `backchannel`
 - `crosstalk`
 
-The `resolve-overlaps` module uses preserved word-level timing to replace detected overlap-group segments with smaller word-run segments when usable timing is available. Resolution expands each overlap window by the configured coalesce gap so nearby same-speaker context can be absorbed into the replacement runs. Groups without usable word timing remain unresolved for later passes or human review.
+The `resolve-overlaps` module uses preserved word-level timing to replace detected overlap-group segments with smaller word-run segments when usable timing is available. Resolution expands each overlap window by the configured coalesce gap so nearby same-speaker context can be absorbed into the replacement runs. Once a segment is selected for replacement, all timed words from that segment participate in word-run construction so text is not clipped at the window boundary. Groups without usable word timing remain unresolved for later passes or human review.
 
 Overlap resolution should be non-destructive. Original segment text, timing, and source metadata must remain recoverable.
 
@@ -389,6 +439,8 @@ A valid merged transcript should satisfy:
 - Every referenced segment exists.
 - Output validates against the selected output schema.
 
+For full-schema trim output, overlap groups are recomputed from retained segments so overlap annotations and group references remain internally consistent after projection.
+
 ## Determinism Requirements
 
 Given the same inputs, config, and application version, `seriatim` should produce byte-stable JSON output where practical.
@@ -403,6 +455,19 @@ To support this:
 - Record application version in output metadata.
 - Record enabled module names and module order in output metadata or report data.
 
+Trim-specific determinism requirements:
+
+- Selector normalization and retained IDs are deterministic.
+- Old-to-new ID mapping in trim reports is emitted in deterministic order.
+- Full-schema overlap recomputation is deterministic for the same input artifact and selector.
+
+Normalize-specific determinism requirements:
+
+- Input-shape detection is deterministic.
+- Segment ordering is deterministic for identical input data.
+- Output IDs are always reassigned sequentially after deterministic sorting.
+- Normalize diagnostic reports are deterministic for identical inputs and configuration.
+
 ## Go Package Layout
 
 ```text
@@ -411,6 +476,8 @@ internal/config/         CLI/env/config loading and validation
 internal/pipeline/       Pipeline orchestration and module registry
 internal/builtin/        Built-in pipeline modules
 internal/artifact/       Conversion from internal model to public output schema
+internal/normalize/      Normalize input parsing, validation, deterministic sorting, schema conversion, and diagnostics
+internal/trim/           Artifact parsing, trim selection, schema conversion, overlap recomputation for full schema
 internal/buildinfo/      Build-time version metadata
 internal/speaker/        Speaker map parsing and lookup
 internal/model/          Canonical and merged transcript models
@@ -422,6 +489,18 @@ schema/                  Public output contract and JSON Schema validation
 
 Package boundaries should follow data ownership. Shared models belong in `internal/model`; stage-specific behavior belongs in the relevant stage package.
 
+For trim:
+
+- `internal/trim` contains pure transformation logic over artifact structs.
+- CLI command code handles only flag parsing, file I/O, and report emission.
+- Transform logic is deterministic and pure except for command-layer I/O.
+
+For normalize:
+
+- `internal/normalize` contains parsing/validation and deterministic schema conversion logic.
+- CLI command code handles flag parsing and delegates execution.
+- Normalize remains artifact-level and does not compose merge pipeline modules.
+
 ## Default Modules
 
 The default pipeline is equivalent to explicit module lists.

internal/artifact/transcript.go

@@ -57,12 +57,12 @@ func FromMerged(cfg config.Config, merged model.MergedTranscript) schema.Transcr
 	}
 }
 
-// DefaultFromMerged converts the internal merged transcript model into the
-// compact default public serialized output contract.
-func DefaultFromMerged(cfg config.Config, merged model.MergedTranscript) schema.DefaultTranscript {
-	segments := make([]schema.DefaultSegment, len(merged.Segments))
+// IntermediateFromMerged converts the internal merged transcript model into the
+// compact intermediate public serialized output contract.
+func IntermediateFromMerged(cfg config.Config, merged model.MergedTranscript) schema.IntermediateTranscript {
+	segments := make([]schema.IntermediateSegment, len(merged.Segments))
 	for index, segment := range merged.Segments {
-		segments[index] = schema.DefaultSegment{
+		segments[index] = schema.IntermediateSegment{
 			ID:         segment.ID,
 			Start:      segment.Start,
 			End:        segment.End,
@@ -72,11 +72,11 @@ func DefaultFromMerged(cfg config.Config, merged model.MergedTranscript) schema.
 		}
 	}
 
-	return schema.DefaultTranscript{
-		Metadata: schema.DefaultMetadata{
+	return schema.IntermediateTranscript{
+		Metadata: schema.IntermediateMetadata{
 			Application:  ApplicationName,
 			Version:      buildinfo.Version,
-			OutputSchema: config.OutputSchemaDefault,
+			OutputSchema: config.OutputSchemaIntermediate,
 		},
 		Segments: segments,
 	}
@@ -110,14 +110,14 @@ func MinimalFromMerged(cfg config.Config, merged model.MergedTranscript) schema.
 // runtime-selected public output contract.
 func SelectedFromMerged(cfg config.Config, merged model.MergedTranscript) any {
 	switch cfg.OutputSchema {
-	case config.OutputSchemaDefault:
-		return DefaultFromMerged(cfg, merged)
 	case config.OutputSchemaMinimal:
 		return MinimalFromMerged(cfg, merged)
-	case config.OutputSchemaSeriatim:
+	case config.OutputSchemaFull:
 		return FromMerged(cfg, merged)
+	case config.OutputSchemaIntermediate:
+		return IntermediateFromMerged(cfg, merged)
 	default:
-		return FromMerged(cfg, merged)
+		return IntermediateFromMerged(cfg, merged)
 	}
 }
 

internal/artifact/transcript_test.go

@@ -23,28 +23,28 @@ func TestFromMergedUsesBuildVersion(t *testing.T) {
 	}
 }
 
-func TestSelectedFromMergedDefaultsToSeriatimTranscript(t *testing.T) {
+func TestSelectedFromMergedDefaultsToIntermediateTranscript(t *testing.T) {
 	got := SelectedFromMerged(config.Config{}, model.MergedTranscript{})
+	if _, ok := got.(schema.IntermediateTranscript); !ok {
+		t.Fatalf("selected artifact type = %T, want schema.IntermediateTranscript", got)
+	}
+}
+
+func TestSelectedFromMergedUsesIntermediateWhenConfigured(t *testing.T) {
+	got := SelectedFromMerged(config.Config{OutputSchema: config.OutputSchemaIntermediate}, model.MergedTranscript{})
+	if _, ok := got.(schema.IntermediateTranscript); !ok {
+		t.Fatalf("selected artifact type = %T, want schema.IntermediateTranscript", got)
+	}
+}
+
+func TestSelectedFromMergedUsesFullWhenConfigured(t *testing.T) {
+	got := SelectedFromMerged(config.Config{OutputSchema: config.OutputSchemaFull}, model.MergedTranscript{})
 	if _, ok := got.(schema.Transcript); !ok {
 		t.Fatalf("selected artifact type = %T, want schema.Transcript", got)
 	}
 }
 
-func TestSelectedFromMergedUsesDefaultWhenConfigured(t *testing.T) {
-	got := SelectedFromMerged(config.Config{OutputSchema: config.OutputSchemaDefault}, model.MergedTranscript{})
-	if _, ok := got.(schema.DefaultTranscript); !ok {
-		t.Fatalf("selected artifact type = %T, want schema.DefaultTranscript", got)
-	}
-}
-
-func TestSelectedFromMergedUsesSeriatimWhenConfigured(t *testing.T) {
-	got := SelectedFromMerged(config.Config{OutputSchema: config.OutputSchemaSeriatim}, model.MergedTranscript{})
-	if _, ok := got.(schema.Transcript); !ok {
-		t.Fatalf("selected artifact type = %T, want schema.Transcript", got)
-	}
-}
-
-func TestDefaultFromMergedEmitsOnlyDefaultShape(t *testing.T) {
+func TestIntermediateFromMergedEmitsOnlyIntermediateShape(t *testing.T) {
 	merged := model.MergedTranscript{
 		Segments: []model.Segment{
 			{
@@ -65,19 +65,19 @@ func TestDefaultFromMergedEmitsOnlyDefaultShape(t *testing.T) {
 		},
 	}
 
-	got := DefaultFromMerged(config.Config{OutputSchema: config.OutputSchemaDefault}, merged)
-	want := schema.DefaultTranscript{
-		Metadata: schema.DefaultMetadata{
+	got := IntermediateFromMerged(config.Config{OutputSchema: config.OutputSchemaIntermediate}, merged)
+	want := schema.IntermediateTranscript{
+		Metadata: schema.IntermediateMetadata{
 			Application:  ApplicationName,
 			Version:      buildinfo.Version,
-			OutputSchema: config.OutputSchemaDefault,
+			OutputSchema: config.OutputSchemaIntermediate,
 		},
-		Segments: []schema.DefaultSegment{
+		Segments: []schema.IntermediateSegment{
 			{ID: 1, Start: 1, End: 2, Speaker: "Alice", Text: "hello", Categories: []string{"backchannel"}},
 		},
 	}
 	if !reflect.DeepEqual(got, want) {
-		t.Fatalf("default transcript = %#v, want %#v", got, want)
+		t.Fatalf("intermediate transcript = %#v, want %#v", got, want)
 	}
 }
 

internal/builtin/postprocess.go

@@ -51,8 +51,8 @@ func (validateOutput) Process(ctx context.Context, in model.MergedTranscript, cf
 	selected := artifact.SelectedFromMerged(cfg, in)
 	var err error
 	switch transcript := selected.(type) {
-	case schema.DefaultTranscript:
-		err = schema.ValidateDefaultTranscript(transcript)
+	case schema.IntermediateTranscript:
+		err = schema.ValidateIntermediateTranscript(transcript)
 	case schema.MinimalTranscript:
 		err = schema.ValidateMinimalTranscript(transcript)
 	case schema.Transcript:

internal/builtin/postprocess_test.go

@@ -81,7 +81,7 @@ func TestValidateOutputUsesMinimalSchemaWhenConfigured(t *testing.T) {
 	}
 }
 
-func TestValidateOutputUsesSeriatimSchemaWhenConfigured(t *testing.T) {
+func TestValidateOutputUsesFullSchemaWhenConfigured(t *testing.T) {
 	merged := model.MergedTranscript{
 		Segments: []model.Segment{
 			{
@@ -100,7 +100,7 @@ func TestValidateOutputUsesSeriatimSchemaWhenConfigured(t *testing.T) {
 	}
 
 	cfg := testConfig()
-	cfg.OutputSchema = config.OutputSchemaSeriatim
+	cfg.OutputSchema = config.OutputSchemaFull
 	got, events, err := validateOutput{}.Process(context.Background(), merged, cfg)
 	if err != nil {
 		t.Fatalf("validate output: %v", err)

internal/cli/merge.go

@@ -15,7 +15,12 @@ func newMergeCommand() *cobra.Command {
 		Use:   "merge",
 		Short: "Run the transcript merge pipeline",
 		RunE: func(cmd *cobra.Command, args []string) error {
-			cfg, err := config.NewMergeConfig(opts)
+			mergeOpts := opts
+			if !cmd.Flags().Changed("output-schema") {
+				mergeOpts.OutputSchema = ""
+			}
+
+			cfg, err := config.NewMergeConfig(mergeOpts)
 			if err != nil {
 				return err
 			}
@@ -32,7 +37,7 @@ func newMergeCommand() *cobra.Command {
 	flags.StringVar(&opts.AutocorrectFile, "autocorrect", "", "autocorrect rules file")
 	flags.StringVar(&opts.InputReader, "input-reader", config.DefaultInputReader, "input reader module")
 	flags.StringVar(&opts.OutputModules, "output-modules", config.DefaultOutputModules, "comma-separated output modules")
-	flags.StringVar(&opts.OutputSchema, "output-schema", config.DefaultOutputSchema, "output JSON schema: default, minimal, or seriatim")
+	flags.StringVar(&opts.OutputSchema, "output-schema", config.DefaultOutputSchema, "output JSON schema: seriatim-minimal, seriatim-intermediate (default), or seriatim-full")
 	flags.StringVar(&opts.PreprocessingModules, "preprocessing-modules", config.DefaultPreprocessingModules, "comma-separated preprocessing modules")
 	flags.StringVar(&opts.PostprocessingModules, "postprocessing-modules", config.DefaultPostprocessingModules, "comma-separated postprocessing modules")
 	flags.StringVar(&opts.CoalesceGap, "coalesce-gap", config.DefaultCoalesceGapValue, "maximum same-speaker gap in seconds for coalesce")

internal/cli/merge_test.go

@@ -114,7 +114,8 @@ func TestMergeWritesMergedOutputAndReport(t *testing.T) {
 	}
 }
 
-func TestMergeWritesDefaultOutputSchema(t *testing.T) {
+func TestMergeWritesIntermediateOutputSchema(t *testing.T) {
+	t.Setenv(config.OutputSchemaEnv, "")
 	dir := t.TempDir()
 	input := writeJSONFile(t, dir, "input.json", `{
 		"segments": [
@@ -131,10 +132,10 @@ func TestMergeWritesDefaultOutputSchema(t *testing.T) {
 		t.Fatalf("merge failed: %v", err)
 	}
 
-	var transcript schema.DefaultTranscript
+	var transcript schema.IntermediateTranscript
 	readJSON(t, output, &transcript)
-	if transcript.Metadata.OutputSchema != config.OutputSchemaDefault {
-		t.Fatalf("output_schema = %q, want default", transcript.Metadata.OutputSchema)
+	if transcript.Metadata.OutputSchema != config.OutputSchemaIntermediate {
+		t.Fatalf("output_schema = %q, want %q", transcript.Metadata.OutputSchema, config.OutputSchemaIntermediate)
 	}
 	if len(transcript.Segments) != 1 {
 		t.Fatalf("segment count = %d, want 1", len(transcript.Segments))
@@ -149,12 +150,12 @@ func TestMergeWritesDefaultOutputSchema(t *testing.T) {
 	}
 	for _, forbidden := range []string{"overlap_groups", "source", "derived_from", "words"} {
 		if strings.Contains(string(outputBytes), forbidden) {
-			t.Fatalf("default output contains %q:\n%s", forbidden, outputBytes)
+			t.Fatalf("intermediate output contains %q:\n%s", forbidden, outputBytes)
 		}
 	}
 }
 
-func TestMergeWritesSeriatimOutputSchema(t *testing.T) {
+func TestMergeWritesFullOutputSchema(t *testing.T) {
 	dir := t.TempDir()
 	input := writeJSONFile(t, dir, "input.json", `{
 		"segments": [
@@ -166,7 +167,7 @@ func TestMergeWritesSeriatimOutputSchema(t *testing.T) {
 	err := executeMergeRaw(
 		"--input-file", input,
 		"--output-file", output,
-		"--output-schema", "seriatim",
+		"--output-schema", config.OutputSchemaFull,
 	)
 	if err != nil {
 		t.Fatalf("merge failed: %v", err)
@@ -203,7 +204,7 @@ func TestMergeWritesMinimalOutputSchema(t *testing.T) {
 	err := executeMerge(
 		"--input-file", input,
 		"--output-file", output,
-		"--output-schema", "minimal",
+		"--output-schema", config.OutputSchemaMinimal,
 		"--report-file", reportPath,
 	)
 	if err != nil {
@@ -215,8 +216,8 @@ func TestMergeWritesMinimalOutputSchema(t *testing.T) {
 	if transcript.Metadata.Application != "seriatim" {
 		t.Fatalf("application = %q, want seriatim", transcript.Metadata.Application)
 	}
-	if transcript.Metadata.OutputSchema != "minimal" {
-		t.Fatalf("output_schema = %q, want minimal", transcript.Metadata.OutputSchema)
+	if transcript.Metadata.OutputSchema != config.OutputSchemaMinimal {
+		t.Fatalf("output_schema = %q, want %q", transcript.Metadata.OutputSchema, config.OutputSchemaMinimal)
 	}
 	if got, want := len(transcript.Segments), 2; got != want {
 		t.Fatalf("segment count = %d, want %d", got, want)
@@ -330,7 +331,7 @@ func TestMergeValidateMinimalOutputBeforeAssignIDsFails(t *testing.T) {
 	err := executeMerge(
 		"--input-file", input,
 		"--output-file", output,
-		"--output-schema", "minimal",
+		"--output-schema", config.OutputSchemaMinimal,
 		"--postprocessing-modules", "validate-output,assign-ids",
 	)
 	if err == nil {
@@ -472,16 +473,16 @@ func TestMergeResolvesOverlapGroupsWithWordRuns(t *testing.T) {
 
 	var transcript model.FinalTranscript
 	readJSON(t, output, &transcript)
-	if len(transcript.OverlapGroups) != 0 {
-		t.Fatalf("overlap groups = %#v, want none", transcript.OverlapGroups)
+	if got, want := len(transcript.OverlapGroups), 1; got != want {
+		t.Fatalf("overlap group count = %d, want %d (%#v)", got, want, transcript.OverlapGroups)
 	}
-	if got, want := len(transcript.Segments), 3; got != want {
+	if got, want := len(transcript.Segments), 2; got != want {
 		t.Fatalf("segment count = %d, want %d", got, want)
 	}
 
-	wantTexts := []string{"outside hello there", "bob reply", "later"}
-	wantSpeakers := []string{"Alice", "Bob", "Alice"}
-	wantRefs := []string{"word-run:1:1:1", "word-run:1:2:1", "word-run:1:1:2"}
+	wantTexts := []string{"outside hello there later", "bob reply"}
+	wantSpeakers := []string{"Alice", "Bob"}
+	wantRefs := []string{"word-run:1:1:1", "word-run:1:2:1"}
 	for index, segment := range transcript.Segments {
 		if segment.ID != index+1 {
 			t.Fatalf("segment %d id = %d, want %d", index, segment.ID, index+1)
@@ -498,8 +499,8 @@ func TestMergeResolvesOverlapGroupsWithWordRuns(t *testing.T) {
 		if segment.SourceSegmentIndex != nil {
 			t.Fatalf("segment %d source_segment_index = %d, want nil", index, *segment.SourceSegmentIndex)
 		}
-		if segment.OverlapGroupID != 0 {
-			t.Fatalf("segment %d overlap_group_id = %d, want 0", index, segment.OverlapGroupID)
+		if segment.OverlapGroupID != 1 {
+			t.Fatalf("segment %d overlap_group_id = %d, want 1", index, segment.OverlapGroupID)
 		}
 	}
 	if !equalStrings(transcript.Segments[0].DerivedFrom, []string{inputA + "#0"}) {
@@ -516,7 +517,7 @@ func TestMergeResolvesOverlapGroupsWithWordRuns(t *testing.T) {
 
 	var rpt report.Report
 	readJSON(t, reportPath, &rpt)
-	if !hasReportEvent(rpt, "postprocessing", "resolve-overlaps", "processed 1 overlap group(s); changed 1; removed 2 original segment(s); created 3 replacement segment(s)") {
+	if !hasReportEvent(rpt, "postprocessing", "resolve-overlaps", "processed 1 overlap group(s); changed 1; removed 2 original segment(s); created 2 replacement segment(s)") {
 		t.Fatal("expected resolve-overlaps summary report event")
 	}
 }
@@ -1927,7 +1928,7 @@ func TestMergeResolveOverlapsPreservesAbsorbedContextPrefix(t *testing.T) {
 		"--input-file", inputA,
 		"--input-file", inputB,
 		"--speakers", speakers,
-		"--output-schema", "minimal",
+		"--output-schema", config.OutputSchemaMinimal,
 		"--output-file", output,
 	)
 	if err != nil {
@@ -1994,7 +1995,7 @@ func executeMerge(args ...string) error {
 	if !hasOutputSchemaFlag(args) {
 		// Most integration tests were written against the full envelope; keep
 		// that behavior unless the caller explicitly asks for another schema.
-		args = append(args, "--output-schema", config.OutputSchemaSeriatim)
+		args = append(args, "--output-schema", config.OutputSchemaFull)
 	}
 	return executeMergeRaw(args...)
 }

internal/cli/normalize.go

@@ -0,0 +1,39 @@
+package cli
+
+import (
+	"github.com/spf13/cobra"
+
+	"gitea.maximumdirect.net/eric/seriatim/internal/config"
+	"gitea.maximumdirect.net/eric/seriatim/internal/normalize"
+)
+
+func newNormalizeCommand() *cobra.Command {
+	var opts config.NormalizeOptions
+
+	cmd := &cobra.Command{
+		Use:   "normalize",
+		Short: "Normalize a transcript artifact into a standard seriatim output shape",
+		RunE: func(cmd *cobra.Command, args []string) error {
+			normalizeOpts := opts
+			if !cmd.Flags().Changed("output-schema") {
+				normalizeOpts.OutputSchema = ""
+			}
+
+			cfg, err := config.NewNormalizeConfig(normalizeOpts)
+			if err != nil {
+				return err
+			}
+
+			return normalize.Run(cmd.Context(), cfg)
+		},
+	}
+
+	flags := cmd.Flags()
+	flags.StringVar(&opts.InputFile, "input-file", "", "input transcript JSON file")
+	flags.StringVar(&opts.OutputFile, "output-file", "", "output transcript JSON file")
+	flags.StringVar(&opts.ReportFile, "report-file", "", "optional report JSON file")
+	flags.StringVar(&opts.OutputSchema, "output-schema", config.DefaultOutputSchema, "output JSON schema: seriatim-minimal, seriatim-intermediate, or seriatim-full")
+	flags.StringVar(&opts.OutputModules, "output-modules", config.DefaultOutputModules, "comma-separated output modules")
+
+	return cmd
+}

internal/cli/normalize_test.go

@@ -0,0 +1,457 @@
+package cli
+
+import (
+	"encoding/json"
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+
+	"gitea.maximumdirect.net/eric/seriatim/internal/config"
+	"gitea.maximumdirect.net/eric/seriatim/internal/report"
+	"gitea.maximumdirect.net/eric/seriatim/schema"
+)
+
+func TestNormalizeCommandIsRecognized(t *testing.T) {
+	cmd := NewRootCommand()
+	cmd.SetArgs([]string{"normalize", "--help"})
+	if err := cmd.Execute(); err != nil {
+		t.Fatalf("normalize command should be recognized: %v", err)
+	}
+}
+
+func TestNormalizeMissingInputFileFails(t *testing.T) {
+	dir := t.TempDir()
+	output := filepath.Join(dir, "normalized.json")
+
+	err := executeNormalize(
+		"--output-file", output,
+	)
+	if err == nil {
+		t.Fatal("expected missing input-file error")
+	}
+	if !strings.Contains(err.Error(), "--input-file is required") {
+		t.Fatalf("unexpected error: %v", err)
+	}
+}
+
+func TestNormalizeMissingOutputFileFails(t *testing.T) {
+	dir := t.TempDir()
+	input := writeJSONFile(t, dir, "input.json", `{"segments":[]}`)
+
+	err := executeNormalize(
+		"--input-file", input,
+	)
+	if err == nil {
+		t.Fatal("expected missing output-file error")
+	}
+	if !strings.Contains(err.Error(), "--output-file is required") {
+		t.Fatalf("unexpected error: %v", err)
+	}
+}
+
+func TestNormalizeInvalidOutputSchemaFails(t *testing.T) {
+	dir := t.TempDir()
+	input := writeJSONFile(t, dir, "input.json", `{"segments":[]}`)
+	output := filepath.Join(dir, "normalized.json")
+
+	err := executeNormalize(
+		"--input-file", input,
+		"--output-file", output,
+		"--output-schema", "compact",
+	)
+	if err == nil {
+		t.Fatal("expected invalid output schema error")
+	}
+	if !strings.Contains(err.Error(), "--output-schema must be one of") {
+		t.Fatalf("unexpected error: %v", err)
+	}
+}
+
+func TestNormalizeInvalidOutputModuleFails(t *testing.T) {
+	dir := t.TempDir()
+	input := writeJSONFile(t, dir, "input.json", `{"segments":[]}`)
+	output := filepath.Join(dir, "normalized.json")
+
+	err := executeNormalize(
+		"--input-file", input,
+		"--output-file", output,
+		"--output-modules", "yaml",
+	)
+	if err == nil {
+		t.Fatal("expected invalid output module error")
+	}
+	if !strings.Contains(err.Error(), "unknown output module") {
+		t.Fatalf("unexpected error: %v", err)
+	}
+}
+
+func TestNormalizeDefaultOutputSchemaIsIntermediate(t *testing.T) {
+	dir := t.TempDir()
+	input := writeJSONFile(t, dir, "input.json", `{
+		"segments": [
+			{"id": 99, "start": 5, "end": 6, "speaker": "Bob", "text": "second", "categories": ["filler"]},
+			{"id": 10, "start": 1, "end": 2, "speaker": "Alice", "text": "first", "categories": ["backchannel"]}
+		]
+	}`)
+	output := filepath.Join(dir, "normalized.json")
+
+	err := executeNormalize(
+		"--input-file", input,
+		"--output-file", output,
+	)
+	if err != nil {
+		t.Fatalf("normalize failed: %v", err)
+	}
+
+	var transcript schema.IntermediateTranscript
+	readJSON(t, output, &transcript)
+	if transcript.Metadata.OutputSchema != config.OutputSchemaIntermediate {
+		t.Fatalf("output schema = %q, want %q", transcript.Metadata.OutputSchema, config.OutputSchemaIntermediate)
+	}
+	if len(transcript.Segments) != 2 {
+		t.Fatalf("segment count = %d, want 2", len(transcript.Segments))
+	}
+	if transcript.Segments[0].ID != 1 || transcript.Segments[1].ID != 2 {
+		t.Fatalf("segment IDs = %d,%d, want 1,2", transcript.Segments[0].ID, transcript.Segments[1].ID)
+	}
+	if transcript.Segments[0].Text != "first" || transcript.Segments[1].Text != "second" {
+		t.Fatalf("unexpected sort order: %#v", transcript.Segments)
+	}
+	if len(transcript.Segments[0].Categories) != 1 || transcript.Segments[0].Categories[0] != "backchannel" {
+		t.Fatalf("expected categories preserved on first segment, got %#v", transcript.Segments[0].Categories)
+	}
+}
+
+func TestNormalizeBareArrayInputToIntermediateOutput(t *testing.T) {
+	dir := t.TempDir()
+	input := writeJSONFile(t, dir, "input.json", `[
+		{"start": 2, "end": 3, "speaker": "Bob", "text": "second"},
+		{"start": 1, "end": 2, "speaker": "Alice", "text": "first"}
+	]`)
+	output := filepath.Join(dir, "normalized.json")
+
+	err := executeNormalize(
+		"--input-file", input,
+		"--output-file", output,
+		"--output-schema", config.OutputSchemaIntermediate,
+	)
+	if err != nil {
+		t.Fatalf("normalize failed: %v", err)
+	}
+
+	var transcript schema.IntermediateTranscript
+	readJSON(t, output, &transcript)
+	if len(transcript.Segments) != 2 {
+		t.Fatalf("segment count = %d, want 2", len(transcript.Segments))
+	}
+	if transcript.Segments[0].Speaker != "Alice" || transcript.Segments[1].Speaker != "Bob" {
+		t.Fatalf("unexpected sorted speakers: %#v", transcript.Segments)
+	}
+}
+
+func TestNormalizeInputIndexTieBreakerIsDeterministic(t *testing.T) {
+	dir := t.TempDir()
+	input := writeJSONFile(t, dir, "input.json", `[
+		{"start": 1, "end": 2, "speaker": "Zulu", "text": "first in"},
+		{"start": 1, "end": 2, "speaker": "Alpha", "text": "second in"}
+	]`)
+	output := filepath.Join(dir, "normalized.json")
+
+	err := executeNormalize(
+		"--input-file", input,
+		"--output-file", output,
+	)
+	if err != nil {
+		t.Fatalf("normalize failed: %v", err)
+	}
+
+	var transcript schema.IntermediateTranscript
+	readJSON(t, output, &transcript)
+	if transcript.Segments[0].Speaker != "Zulu" || transcript.Segments[1].Speaker != "Alpha" {
+		t.Fatalf("tie-break order mismatch: %#v", transcript.Segments)
+	}
+}
+
+func TestNormalizeMinimalSchemaOmitsCategories(t *testing.T) {
+	dir := t.TempDir()
+	input := writeJSONFile(t, dir, "input.json", `{
+		"segments": [
+			{"start": 1, "end": 2, "speaker": "Alice", "text": "first", "categories": ["filler"]}
+		]
+	}`)
+	output := filepath.Join(dir, "normalized.json")
+
+	err := executeNormalize(
+		"--input-file", input,
+		"--output-file", output,
+		"--output-schema", config.OutputSchemaMinimal,
+	)
+	if err != nil {
+		t.Fatalf("normalize failed: %v", err)
+	}
+
+	var transcript schema.MinimalTranscript
+	readJSON(t, output, &transcript)
+	if transcript.Metadata.OutputSchema != config.OutputSchemaMinimal {
+		t.Fatalf("output schema = %q, want %q", transcript.Metadata.OutputSchema, config.OutputSchemaMinimal)
+	}
+	if len(transcript.Segments) != 1 || transcript.Segments[0].ID != 1 {
+		t.Fatalf("unexpected minimal output: %#v", transcript.Segments)
+	}
+	bytes, readErr := os.ReadFile(output)
+	if readErr != nil {
+		t.Fatalf("read output: %v", readErr)
+	}
+	if strings.Contains(string(bytes), "categories") {
+		t.Fatalf("minimal output unexpectedly contains categories:\n%s", string(bytes))
+	}
+}
+
+func TestNormalizeFullSchemaOutputValidatesAndHasProvenanceFallback(t *testing.T) {
+	dir := t.TempDir()
+	input := writeJSONFile(t, dir, "input.json", `[
+		{"start": 1, "end": 2, "speaker": "Alice", "text": "first"},
+		{"start": 3, "end": 4, "speaker": "Bob", "text": "second", "source":"custom.json", "source_segment_index": 7}
+	]`)
+	output := filepath.Join(dir, "normalized.json")
+
+	err := executeNormalize(
+		"--input-file", input,
+		"--output-file", output,
+		"--output-schema", config.OutputSchemaFull,
+	)
+	if err != nil {
+		t.Fatalf("normalize failed: %v", err)
+	}
+
+	var transcript schema.Transcript
+	readJSON(t, output, &transcript)
+	if err := schema.ValidateTranscript(transcript); err != nil {
+		t.Fatalf("full output should validate: %v", err)
+	}
+	if len(transcript.Segments) != 2 {
+		t.Fatalf("segment count = %d, want 2", len(transcript.Segments))
+	}
+	if transcript.Segments[0].Source != filepath.Base(input) {
+		t.Fatalf("source fallback = %q, want %q", transcript.Segments[0].Source, filepath.Base(input))
+	}
+	if transcript.Segments[0].SourceSegmentIndex == nil || *transcript.Segments[0].SourceSegmentIndex != 0 {
+		t.Fatalf("source_segment_index fallback = %v, want 0", transcript.Segments[0].SourceSegmentIndex)
+	}
+	if transcript.Segments[1].Source != "custom.json" {
+		t.Fatalf("explicit source preserved = %q, want custom.json", transcript.Segments[1].Source)
+	}
+	if transcript.Segments[1].SourceSegmentIndex == nil || *transcript.Segments[1].SourceSegmentIndex != 7 {
+		t.Fatalf("explicit source_segment_index preserved = %v, want 7", transcript.Segments[1].SourceSegmentIndex)
+	}
+	if transcript.OverlapGroups == nil || len(transcript.OverlapGroups) != 0 {
+		t.Fatalf("overlap_groups = %#v, want empty array", transcript.OverlapGroups)
+	}
+}
+
+func TestNormalizeEmptySegmentsArrayProducesValidOutput(t *testing.T) {
+	dir := t.TempDir()
+	input := writeJSONFile(t, dir, "input.json", `{"segments":[]}`)
+	output := filepath.Join(dir, "normalized.json")
+
+	err := executeNormalize(
+		"--input-file", input,
+		"--output-file", output,
+	)
+	if err != nil {
+		t.Fatalf("normalize failed: %v", err)
+	}
+
+	var transcript schema.IntermediateTranscript
+	readJSON(t, output, &transcript)
+	if len(transcript.Segments) != 0 {
+		t.Fatalf("segment count = %d, want 0", len(transcript.Segments))
+	}
+	if err := schema.ValidateIntermediateTranscript(transcript); err != nil {
+		t.Fatalf("intermediate output should validate: %v", err)
+	}
+}
+
+func TestNormalizeSelectedOutputSchemaIsHonored(t *testing.T) {
+	dir := t.TempDir()
+	input := writeJSONFile(t, dir, "input.json", `{"segments":[{"start":1,"end":2,"speaker":"A","text":"one"}]}`)
+	output := filepath.Join(dir, "normalized.json")
+
+	err := executeNormalize(
+		"--input-file", input,
+		"--output-file", output,
+		"--output-schema", config.OutputSchemaMinimal,
+	)
+	if err != nil {
+		t.Fatalf("normalize failed: %v", err)
+	}
+
+	var transcript schema.MinimalTranscript
+	readJSON(t, output, &transcript)
+	if transcript.Metadata.OutputSchema != config.OutputSchemaMinimal {
+		t.Fatalf("output schema = %q, want %q", transcript.Metadata.OutputSchema, config.OutputSchemaMinimal)
+	}
+}
+
+func TestNormalizeReportFileWrittenAndContainsObjectInputShape(t *testing.T) {
+	dir := t.TempDir()
+	input := writeJSONFile(t, dir, "input.json", `{"segments":[{"start":1,"end":2,"speaker":"A","text":"one"}]}`)
+	output := filepath.Join(dir, "normalized.json")
+	reportPath := filepath.Join(dir, "report.json")
+
+	err := executeNormalize(
+		"--input-file", input,
+		"--output-file", output,
+		"--report-file", reportPath,
+	)
+	if err != nil {
+		t.Fatalf("normalize failed: %v", err)
+	}
+
+	var rpt report.Report
+	readJSON(t, reportPath, &rpt)
+	audit := extractNormalizeAudit(t, rpt)
+	if audit.InputShape != "object_with_segments" {
+		t.Fatalf("input shape = %q, want object_with_segments", audit.InputShape)
+	}
+	if audit.InputSegmentCount != 1 {
+		t.Fatalf("input segment count = %d, want 1", audit.InputSegmentCount)
+	}
+	if audit.OutputSchema != config.OutputSchemaIntermediate {
+		t.Fatalf("output schema = %q, want %q", audit.OutputSchema, config.OutputSchemaIntermediate)
+	}
+	if len(audit.OutputModules) != 1 || audit.OutputModules[0] != "json" {
+		t.Fatalf("output modules = %v, want [json]", audit.OutputModules)
+	}
+}
+
+func TestNormalizeReportIncludesBareArrayShape(t *testing.T) {
+	dir := t.TempDir()
+	input := writeJSONFile(t, dir, "input.json", `[{"start":1,"end":2,"speaker":"A","text":"one"}]`)
+	output := filepath.Join(dir, "normalized.json")
+	reportPath := filepath.Join(dir, "report.json")
+
+	err := executeNormalize(
+		"--input-file", input,
+		"--output-file", output,
+		"--report-file", reportPath,
+	)
+	if err != nil {
+		t.Fatalf("normalize failed: %v", err)
+	}
+
+	var rpt report.Report
+	readJSON(t, reportPath, &rpt)
+	audit := extractNormalizeAudit(t, rpt)
+	if audit.InputShape != "bare_segments_array" {
+		t.Fatalf("input shape = %q, want bare_segments_array", audit.InputShape)
+	}
+}
+
+func TestNormalizeReportDoesNotIncludeTranscriptText(t *testing.T) {
+	dir := t.TempDir()
+	const segmentText = "normalize-report-secret-text"
+	input := writeJSONFile(t, dir, "input.json", `[{"start":1,"end":2,"speaker":"A","text":"`+segmentText+`"}]`)
+	output := filepath.Join(dir, "normalized.json")
+	reportPath := filepath.Join(dir, "report.json")
+
+	err := executeNormalize(
+		"--input-file", input,
+		"--output-file", output,
+		"--report-file", reportPath,
+	)
+	if err != nil {
+		t.Fatalf("normalize failed: %v", err)
+	}
+
+	var rpt report.Report
+	readJSON(t, reportPath, &rpt)
+	for _, event := range rpt.Events {
+		if strings.Contains(event.Message, segmentText) {
+			t.Fatalf("report unexpectedly contained transcript text in event %#v", event)
+		}
+	}
+}
+
+func TestNormalizeReportEmptyInputEmitsWarning(t *testing.T) {
+	dir := t.TempDir()
+	input := writeJSONFile(t, dir, "input.json", `{"segments":[]}`)
+	output := filepath.Join(dir, "normalized.json")
+	reportPath := filepath.Join(dir, "report.json")
+
+	err := executeNormalize(
+		"--input-file", input,
+		"--output-file", output,
+		"--report-file", reportPath,
+	)
+	if err != nil {
+		t.Fatalf("normalize failed: %v", err)
+	}
+
+	var rpt report.Report
+	readJSON(t, reportPath, &rpt)
+	found := false
+	for _, event := range rpt.Events {
+		if event.Stage == "normalize" && event.Module == "normalize" && event.Severity == report.SeverityWarning &&
+			strings.Contains(event.Message, "zero segments") {
+			found = true
+			break
+		}
+	}
+	if !found {
+		t.Fatalf("expected empty transcript warning event, got %#v", rpt.Events)
+	}
+}
+
+func TestNormalizeReportWriteFailureReturnsClearError(t *testing.T) {
+	dir := t.TempDir()
+	input := writeJSONFile(t, dir, "input.json", `{"segments":[{"start":1,"end":2,"speaker":"A","text":"one"}]}`)
+	output := filepath.Join(dir, "normalized.json")
+
+	err := executeNormalize(
+		"--input-file", input,
+		"--output-file", output,
+		"--report-file", dir,
+	)
+	if err == nil {
+		t.Fatal("expected report write failure")
+	}
+	if !strings.Contains(err.Error(), "write --report-file") {
+		t.Fatalf("unexpected error: %v", err)
+	}
+}
+
+func executeNormalize(args ...string) error {
+	cmd := NewRootCommand()
+	cmd.SetArgs(append([]string{"normalize"}, args...))
+	return cmd.Execute()
+}
+
+type normalizeAudit struct {
+	Command                string   `json:"command"`
+	InputFile              string   `json:"input_file"`
+	OutputFile             string   `json:"output_file"`
+	InputShape             string   `json:"input_shape"`
+	InputSegmentCount      int      `json:"input_segment_count"`
+	OutputSchema           string   `json:"output_schema"`
+	OutputModules          []string `json:"output_modules"`
+	IDsReassigned          bool     `json:"ids_reassigned"`
+	SortingChangedInput    bool     `json:"sorting_changed_input_order"`
+	SegmentsWithCategories int      `json:"segments_with_categories"`
+}
+
+func extractNormalizeAudit(t *testing.T, rpt report.Report) normalizeAudit {
+	t.Helper()
+	for _, event := range rpt.Events {
+		if event.Stage == "normalize" && event.Module == "normalize-audit" {
+			var audit normalizeAudit
+			if err := json.Unmarshal([]byte(event.Message), &audit); err != nil {
+				t.Fatalf("decode normalize audit: %v", err)
+			}
+			return audit
+		}
+	}
+	t.Fatalf("missing normalize-audit event: %#v", rpt.Events)
+	return normalizeAudit{}
+}

internal/cli/root.go

@@ -10,12 +10,14 @@ import (
 func NewRootCommand() *cobra.Command {
 	cmd := &cobra.Command{
 		Use:           "seriatim",
-		Short:         "Merge per-speaker transcripts into a chronological transcript",
+		Short:         "Merge, trim, and normalize transcript artifacts",
 		Version:       buildinfo.Version,
 		SilenceErrors: true,
 		SilenceUsage:  true,
 	}
 
 	cmd.AddCommand(newMergeCommand())
+	cmd.AddCommand(newNormalizeCommand())
+	cmd.AddCommand(newTrimCommand())
 	return cmd
 }

internal/cli/trim.go

@@ -0,0 +1,191 @@
+package cli
+
+import (
+	"encoding/json"
+	"fmt"
+	"os"
+	"sort"
+
+	"github.com/spf13/cobra"
+
+	"gitea.maximumdirect.net/eric/seriatim/internal/config"
+	"gitea.maximumdirect.net/eric/seriatim/internal/report"
+	triminternal "gitea.maximumdirect.net/eric/seriatim/internal/trim"
+)
+
+type trimAuditReport struct {
+	Operation               string          `json:"operation"`
+	InputFile               string          `json:"input_file"`
+	OutputFile              string          `json:"output_file"`
+	InputSchema             string          `json:"input_schema"`
+	OutputSchema            string          `json:"output_schema"`
+	Mode                    string          `json:"mode"`
+	Selector                string          `json:"selector"`
+	SelectedIDs             []int           `json:"selected_ids"`
+	AllowEmpty              bool            `json:"allow_empty"`
+	InputSegmentCount       int             `json:"input_segment_count"`
+	RetainedSegmentCount    int             `json:"retained_segment_count"`
+	RemovedSegmentCount     int             `json:"removed_segment_count"`
+	RemovedInputIDs         []int           `json:"removed_input_ids"`
+	OldToNewIDMapping       []trimIDMapping `json:"old_to_new_id_mapping"`
+	OverlapGroupsRecomputed bool            `json:"overlap_groups_recomputed"`
+}
+
+type trimIDMapping struct {
+	OldID int `json:"old_id"`
+	NewID int `json:"new_id"`
+}
+
+func newTrimCommand() *cobra.Command {
+	var opts config.TrimOptions
+
+	cmd := &cobra.Command{
+		Use:   "trim",
+		Short: "Trim an existing seriatim transcript artifact by segment ID",
+		RunE: func(cmd *cobra.Command, args []string) error {
+			trimOpts := opts
+			if !cmd.Flags().Changed("output-schema") {
+				trimOpts.OutputSchema = ""
+			}
+
+			cfg, err := config.NewTrimConfig(trimOpts)
+			if err != nil {
+				return err
+			}
+
+			selector, err := triminternal.ParseSelector(cfg.Selector)
+			if err != nil {
+				return fmt.Errorf("invalid selector %q: %w", cfg.Selector, err)
+			}
+
+			data, err := os.ReadFile(cfg.InputFile)
+			if err != nil {
+				return fmt.Errorf("read --input-file %q: %w", cfg.InputFile, err)
+			}
+
+			artifact, err := triminternal.ParseArtifactJSON(data)
+			if err != nil {
+				return fmt.Errorf("--input-file %q: %w", cfg.InputFile, err)
+			}
+			inputSegmentCount := artifact.SegmentCount()
+			inputSchema := artifact.Schema
+
+			mode := triminternal.ModeKeep
+			if cfg.Mode == "remove" {
+				mode = triminternal.ModeRemove
+			}
+
+			trimmed, err := triminternal.ApplyArtifact(artifact, triminternal.Options{
+				Mode:       mode,
+				Selector:   selector,
+				AllowEmpty: cfg.AllowEmpty,
+			})
+			if err != nil {
+				return err
+			}
+
+			outputSchema := artifact.Schema
+			if cfg.OutputSchema != "" {
+				outputSchema = cfg.OutputSchema
+			}
+
+			outputArtifact, err := triminternal.ConvertArtifact(trimmed.Artifact, outputSchema)
+			if err != nil {
+				return err
+			}
+
+			if err := triminternal.ValidateArtifact(outputArtifact); err != nil {
+				return fmt.Errorf("validate trimmed output: %w", err)
+			}
+
+			if err := writeOutputJSON(cfg.OutputFile, outputArtifact.Value()); err != nil {
+				return err
+			}
+
+			if cfg.ReportFile != "" {
+				audit := trimAuditReport{
+					Operation:               "trim",
+					InputFile:               cfg.InputFile,
+					OutputFile:              cfg.OutputFile,
+					InputSchema:             inputSchema,
+					OutputSchema:            outputArtifact.Schema,
+					Mode:                    cfg.Mode,
+					Selector:                cfg.Selector,
+					SelectedIDs:             selector.IDs(),
+					AllowEmpty:              cfg.AllowEmpty,
+					InputSegmentCount:       inputSegmentCount,
+					RetainedSegmentCount:    len(trimmed.OldToNewID),
+					RemovedSegmentCount:     len(trimmed.RemovedIDs),
+					RemovedInputIDs:         append([]int(nil), trimmed.RemovedIDs...),
+					OldToNewIDMapping:       orderedIDMapping(trimmed.OldToNewID),
+					OverlapGroupsRecomputed: trimmed.OverlapGroupsRecomputed,
+				}
+				auditJSON, err := json.Marshal(audit)
+				if err != nil {
+					return fmt.Errorf("marshal trim audit report: %w", err)
+				}
+
+				rpt := report.Report{
+					Metadata: report.Metadata{
+						Application:   outputArtifact.Application(),
+						Version:       outputArtifact.Version(),
+						InputReader:   "trim-artifact",
+						InputFiles:    []string{cfg.InputFile},
+						OutputModules: []string{"json"},
+					},
+					Events: []report.Event{
+						report.Info("trim", "trim", fmt.Sprintf("trimmed %d input segment(s) into %d output segment(s) with mode=%s", inputSegmentCount, outputArtifact.SegmentCount(), cfg.Mode)),
+						report.Info("trim", "trim-audit", string(auditJSON)),
+						report.Info("trim", "validate-output", fmt.Sprintf("validated %d output segment(s)", outputArtifact.SegmentCount())),
+						report.Info("output", "json", "wrote transcript JSON"),
+					},
+				}
+				if err := report.WriteJSON(cfg.ReportFile, rpt); err != nil {
+					return err
+				}
+			}
+
+			return nil
+		},
+	}
+
+	flags := cmd.Flags()
+	flags.StringVar(&opts.InputFile, "input-file", "", "input seriatim transcript artifact JSON file")
+	flags.StringVar(&opts.OutputFile, "output-file", "", "output transcript JSON file")
+	flags.StringVar(&opts.ReportFile, "report-file", "", "optional report JSON file")
+	flags.StringVar(&opts.Keep, "keep", "", "segment ID selector to keep (for example: 1-10,15)")
+	flags.StringVar(&opts.Remove, "remove", "", "segment ID selector to remove (for example: 1-10,15)")
+	flags.StringVar(&opts.OutputSchema, "output-schema", "", "optional output JSON schema override: seriatim-minimal, seriatim-intermediate, or seriatim-full")
+	flags.BoolVar(&opts.AllowEmpty, "allow-empty", false, "allow trimming to an empty transcript")
+
+	return cmd
+}
+
+func writeOutputJSON(path string, value any) error {
+	file, err := os.Create(path)
+	if err != nil {
+		return err
+	}
+	defer file.Close()
+
+	enc := json.NewEncoder(file)
+	enc.SetIndent("", "  ")
+	return enc.Encode(value)
+}
+
+func orderedIDMapping(mapping map[int]int) []trimIDMapping {
+	keys := make([]int, 0, len(mapping))
+	for oldID := range mapping {
+		keys = append(keys, oldID)
+	}
+	sort.Ints(keys)
+
+	pairs := make([]trimIDMapping, 0, len(keys))
+	for _, oldID := range keys {
+		pairs = append(pairs, trimIDMapping{
+			OldID: oldID,
+			NewID: mapping[oldID],
+		})
+	}
+	return pairs
+}

internal/cli/trim_test.go

@@ -0,0 +1,758 @@
+package cli
+
+import (
+	"encoding/json"
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+
+	"gitea.maximumdirect.net/eric/seriatim/internal/config"
+	"gitea.maximumdirect.net/eric/seriatim/internal/report"
+	"gitea.maximumdirect.net/eric/seriatim/schema"
+)
+
+func TestTrimKeepModeEndToEnd(t *testing.T) {
+	dir := t.TempDir()
+	input := writeTrimFullFixture(t, dir, "input.json")
+	output := filepath.Join(dir, "trimmed.json")
+
+	err := executeTrim(
+		"--input-file", input,
+		"--output-file", output,
+		"--keep", "2,4",
+	)
+	if err != nil {
+		t.Fatalf("trim failed: %v", err)
+	}
+
+	var transcript schema.Transcript
+	readJSON(t, output, &transcript)
+	if len(transcript.Segments) != 2 {
+		t.Fatalf("segment count = %d, want 2", len(transcript.Segments))
+	}
+	if transcript.Segments[0].Text != "two" || transcript.Segments[1].Text != "four" {
+		t.Fatalf("unexpected kept text order: %#v", transcript.Segments)
+	}
+	assertSequentialIDs(t, []int{transcript.Segments[0].ID, transcript.Segments[1].ID})
+}
+
+func TestTrimRemoveModeEndToEnd(t *testing.T) {
+	dir := t.TempDir()
+	input := writeTrimFullFixture(t, dir, "input.json")
+	output := filepath.Join(dir, "trimmed.json")
+
+	err := executeTrim(
+		"--input-file", input,
+		"--output-file", output,
+		"--remove", "2,4",
+	)
+	if err != nil {
+		t.Fatalf("trim failed: %v", err)
+	}
+
+	var transcript schema.Transcript
+	readJSON(t, output, &transcript)
+	if len(transcript.Segments) != 2 {
+		t.Fatalf("segment count = %d, want 2", len(transcript.Segments))
+	}
+	if transcript.Segments[0].Text != "one" || transcript.Segments[1].Text != "three" {
+		t.Fatalf("unexpected remaining text order: %#v", transcript.Segments)
+	}
+	assertSequentialIDs(t, []int{transcript.Segments[0].ID, transcript.Segments[1].ID})
+}
+
+func TestTrimMutualExclusionFailure(t *testing.T) {
+	dir := t.TempDir()
+	input := writeTrimFullFixture(t, dir, "input.json")
+	output := filepath.Join(dir, "trimmed.json")
+
+	err := executeTrim(
+		"--input-file", input,
+		"--output-file", output,
+		"--keep", "1",
+		"--remove", "2",
+	)
+	if err == nil {
+		t.Fatal("expected mutual exclusion error")
+	}
+	if !strings.Contains(err.Error(), "mutually exclusive") {
+		t.Fatalf("unexpected error: %v", err)
+	}
+}
+
+func TestTrimMissingSelectionFailure(t *testing.T) {
+	dir := t.TempDir()
+	input := writeTrimFullFixture(t, dir, "input.json")
+	output := filepath.Join(dir, "trimmed.json")
+
+	err := executeTrim(
+		"--input-file", input,
+		"--output-file", output,
+	)
+	if err == nil {
+		t.Fatal("expected selection flag error")
+	}
+	if !strings.Contains(err.Error(), "exactly one of --keep or --remove is required") {
+		t.Fatalf("unexpected error: %v", err)
+	}
+}
+
+func TestTrimInvalidSelectedIDFailure(t *testing.T) {
+	dir := t.TempDir()
+	input := writeTrimFullFixture(t, dir, "input.json")
+	output := filepath.Join(dir, "trimmed.json")
+
+	err := executeTrim(
+		"--input-file", input,
+		"--output-file", output,
+		"--keep", "99",
+	)
+	if err == nil {
+		t.Fatal("expected missing selected ID error")
+	}
+	if !strings.Contains(err.Error(), "does not exist") {
+		t.Fatalf("unexpected error: %v", err)
+	}
+}
+
+func TestTrimOmittedOutputSchemaPreservesInputSchema(t *testing.T) {
+	dir := t.TempDir()
+	input := writeTrimMinimalFixture(t, dir, "input-minimal.json")
+	output := filepath.Join(dir, "trimmed.json")
+
+	err := executeTrim(
+		"--input-file", input,
+		"--output-file", output,
+		"--keep", "1",
+	)
+	if err != nil {
+		t.Fatalf("trim failed: %v", err)
+	}
+
+	var transcript schema.MinimalTranscript
+	readJSON(t, output, &transcript)
+	if transcript.Metadata.OutputSchema != config.OutputSchemaMinimal {
+		t.Fatalf("output_schema = %q, want %q", transcript.Metadata.OutputSchema, config.OutputSchemaMinimal)
+	}
+	if len(transcript.Segments) != 1 || transcript.Segments[0].ID != 1 {
+		t.Fatalf("unexpected minimal trim output: %#v", transcript.Segments)
+	}
+}
+
+func TestTrimExplicitOutputSchemaChangesOutputSchema(t *testing.T) {
+	dir := t.TempDir()
+	input := writeTrimFullFixture(t, dir, "input.json")
+	output := filepath.Join(dir, "trimmed.json")
+
+	err := executeTrim(
+		"--input-file", input,
+		"--output-file", output,
+		"--keep", "1,3",
+		"--output-schema", config.OutputSchemaMinimal,
+	)
+	if err != nil {
+		t.Fatalf("trim failed: %v", err)
+	}
+
+	var transcript schema.MinimalTranscript
+	readJSON(t, output, &transcript)
+	if transcript.Metadata.OutputSchema != config.OutputSchemaMinimal {
+		t.Fatalf("output_schema = %q, want %q", transcript.Metadata.OutputSchema, config.OutputSchemaMinimal)
+	}
+	if len(transcript.Segments) != 2 {
+		t.Fatalf("segment count = %d, want 2", len(transcript.Segments))
+	}
+	assertSequentialIDs(t, []int{transcript.Segments[0].ID, transcript.Segments[1].ID})
+}
+
+func TestTrimExplicitOutputSchemaConvertsMinimalToIntermediate(t *testing.T) {
+	dir := t.TempDir()
+	input := writeTrimMinimalFixture(t, dir, "input-minimal.json")
+	output := filepath.Join(dir, "trimmed.json")
+
+	err := executeTrim(
+		"--input-file", input,
+		"--output-file", output,
+		"--keep", "1-2",
+		"--output-schema", config.OutputSchemaIntermediate,
+	)
+	if err != nil {
+		t.Fatalf("trim failed: %v", err)
+	}
+
+	var transcript schema.IntermediateTranscript
+	readJSON(t, output, &transcript)
+	if transcript.Metadata.OutputSchema != config.OutputSchemaIntermediate {
+		t.Fatalf("output_schema = %q, want %q", transcript.Metadata.OutputSchema, config.OutputSchemaIntermediate)
+	}
+	if len(transcript.Segments) != 2 {
+		t.Fatalf("segment count = %d, want 2", len(transcript.Segments))
+	}
+	assertSequentialIDs(t, []int{transcript.Segments[0].ID, transcript.Segments[1].ID})
+}
+
+func TestTrimIntermediateInputPreservesIntermediateOutputAndCategories(t *testing.T) {
+	dir := t.TempDir()
+	input := writeTrimIntermediateFixture(t, dir, "input-intermediate.json")
+	output := filepath.Join(dir, "trimmed.json")
+
+	err := executeTrim(
+		"--input-file", input,
+		"--output-file", output,
+		"--keep", "2",
+	)
+	if err != nil {
+		t.Fatalf("trim failed: %v", err)
+	}
+
+	var transcript schema.IntermediateTranscript
+	readJSON(t, output, &transcript)
+	if transcript.Metadata.OutputSchema != config.OutputSchemaIntermediate {
+		t.Fatalf("output_schema = %q, want %q", transcript.Metadata.OutputSchema, config.OutputSchemaIntermediate)
+	}
+	if len(transcript.Segments) != 1 {
+		t.Fatalf("segment count = %d, want 1", len(transcript.Segments))
+	}
+	if transcript.Segments[0].ID != 1 {
+		t.Fatalf("segment ID = %d, want 1", transcript.Segments[0].ID)
+	}
+	assertIntSliceEqual(t, []int{len(transcript.Segments[0].Categories)}, []int{2})
+	if transcript.Segments[0].Categories[0] != "filler" || transcript.Segments[0].Categories[1] != "backchannel" {
+		t.Fatalf("categories = %v, want [filler backchannel]", transcript.Segments[0].Categories)
+	}
+}
+
+func TestTrimFullInputPreservesFullShapeAndRecomputesOverlapGroups(t *testing.T) {
+	dir := t.TempDir()
+	input := writeTrimFullOverlapFixture(t, dir, "input-full-overlap.json")
+	output := filepath.Join(dir, "trimmed.json")
+
+	err := executeTrim(
+		"--input-file", input,
+		"--output-file", output,
+		"--keep", "1,2",
+	)
+	if err != nil {
+		t.Fatalf("trim failed: %v", err)
+	}
+
+	var transcript schema.Transcript
+	readJSON(t, output, &transcript)
+	if len(transcript.Segments) != 2 {
+		t.Fatalf("segment count = %d, want 2", len(transcript.Segments))
+	}
+	assertSequentialIDs(t, []int{transcript.Segments[0].ID, transcript.Segments[1].ID})
+	if len(transcript.OverlapGroups) != 1 {
+		t.Fatalf("overlap group count = %d, want 1", len(transcript.OverlapGroups))
+	}
+	if transcript.OverlapGroups[0].ID != 1 {
+		t.Fatalf("overlap group id = %d, want 1", transcript.OverlapGroups[0].ID)
+	}
+	if transcript.Segments[0].OverlapGroupID != 1 || transcript.Segments[1].OverlapGroupID != 1 {
+		t.Fatalf("segment overlap IDs = %d,%d, want 1,1", transcript.Segments[0].OverlapGroupID, transcript.Segments[1].OverlapGroupID)
+	}
+}
+
+func TestTrimMalformedSelectorFailsWithClearError(t *testing.T) {
+	dir := t.TempDir()
+	input := writeTrimFullFixture(t, dir, "input.json")
+	output := filepath.Join(dir, "trimmed.json")
+
+	err := executeTrim(
+		"--input-file", input,
+		"--output-file", output,
+		"--keep", "1-",
+	)
+	if err == nil {
+		t.Fatal("expected malformed selector error")
+	}
+	if !strings.Contains(err.Error(), "invalid selector") || !strings.Contains(err.Error(), "malformed element") {
+		t.Fatalf("unexpected error: %v", err)
+	}
+}
+
+func TestTrimMalformedInputArtifactFailsClearly(t *testing.T) {
+	dir := t.TempDir()
+	input := writeJSONFile(t, dir, "broken.json", `{"metadata":`)
+	output := filepath.Join(dir, "trimmed.json")
+
+	err := executeTrim(
+		"--input-file", input,
+		"--output-file", output,
+		"--keep", "1",
+	)
+	if err == nil {
+		t.Fatal("expected malformed artifact error")
+	}
+	if !strings.Contains(err.Error(), "input JSON is malformed") {
+		t.Fatalf("unexpected error: %v", err)
+	}
+}
+
+func TestTrimDuplicateInputSegmentIDsFail(t *testing.T) {
+	dir := t.TempDir()
+	input := writeTrimMinimalWithIDsFixture(t, dir, "input-dup.json", []int{1, 1})
+	output := filepath.Join(dir, "trimmed.json")
+
+	err := executeTrim(
+		"--input-file", input,
+		"--output-file", output,
+		"--keep", "1",
+	)
+	if err == nil {
+		t.Fatal("expected duplicate segment ID failure")
+	}
+	if !strings.Contains(err.Error(), "not a valid seriatim output artifact") {
+		t.Fatalf("unexpected error: %v", err)
+	}
+}
+
+func TestTrimNonSequentialInputSegmentIDsFail(t *testing.T) {
+	dir := t.TempDir()
+	input := writeTrimMinimalWithIDsFixture(t, dir, "input-nonseq.json", []int{1, 3})
+	output := filepath.Join(dir, "trimmed.json")
+
+	err := executeTrim(
+		"--input-file", input,
+		"--output-file", output,
+		"--keep", "1",
+	)
+	if err == nil {
+		t.Fatal("expected non-sequential segment ID failure")
+	}
+	if !strings.Contains(err.Error(), "not a valid seriatim output artifact") {
+		t.Fatalf("unexpected error: %v", err)
+	}
+}
+
+func TestTrimKeepSelectorWithOverlappingRanges(t *testing.T) {
+	dir := t.TempDir()
+	input := writeTrimFullFixture(t, dir, "input.json")
+	output := filepath.Join(dir, "trimmed.json")
+
+	err := executeTrim(
+		"--input-file", input,
+		"--output-file", output,
+		"--keep", "1-3,2-4",
+	)
+	if err != nil {
+		t.Fatalf("trim failed: %v", err)
+	}
+
+	var transcript schema.Transcript
+	readJSON(t, output, &transcript)
+	if len(transcript.Segments) != 4 {
+		t.Fatalf("segment count = %d, want 4", len(transcript.Segments))
+	}
+	assertSequentialIDs(t, []int{
+		transcript.Segments[0].ID,
+		transcript.Segments[1].ID,
+		transcript.Segments[2].ID,
+		transcript.Segments[3].ID,
+	})
+}
+
+func TestTrimRemoveSelectorWithOverlappingRanges(t *testing.T) {
+	dir := t.TempDir()
+	input := writeTrimFullFixture(t, dir, "input.json")
+	output := filepath.Join(dir, "trimmed.json")
+
+	err := executeTrim(
+		"--input-file", input,
+		"--output-file", output,
+		"--remove", "2-3,3-4",
+	)
+	if err != nil {
+		t.Fatalf("trim failed: %v", err)
+	}
+
+	var transcript schema.Transcript
+	readJSON(t, output, &transcript)
+	if len(transcript.Segments) != 1 {
+		t.Fatalf("segment count = %d, want 1", len(transcript.Segments))
+	}
+	if transcript.Segments[0].Text != "one" {
+		t.Fatalf("remaining segment = %#v, want one", transcript.Segments[0])
+	}
+}
+
+func TestTrimSelectorOrderDoesNotAffectTranscriptOrder(t *testing.T) {
+	dir := t.TempDir()
+	input := writeTrimFullFixture(t, dir, "input.json")
+	output := filepath.Join(dir, "trimmed.json")
+
+	err := executeTrim(
+		"--input-file", input,
+		"--output-file", output,
+		"--keep", "4,1,3",
+	)
+	if err != nil {
+		t.Fatalf("trim failed: %v", err)
+	}
+
+	var transcript schema.Transcript
+	readJSON(t, output, &transcript)
+	if len(transcript.Segments) != 3 {
+		t.Fatalf("segment count = %d, want 3", len(transcript.Segments))
+	}
+	got := []string{
+		transcript.Segments[0].Text,
+		transcript.Segments[1].Text,
+		transcript.Segments[2].Text,
+	}
+	want := []string{"one", "three", "four"}
+	if got[0] != want[0] || got[1] != want[1] || got[2] != want[2] {
+		t.Fatalf("segment text order = %v, want %v", got, want)
+	}
+}
+
+func TestTrimAllowEmptyBehavior(t *testing.T) {
+	dir := t.TempDir()
+	input := writeTrimFullFixture(t, dir, "input.json")
+	output := filepath.Join(dir, "trimmed.json")
+
+	err := executeTrim(
+		"--input-file", input,
+		"--output-file", output,
+		"--remove", "1-4",
+	)
+	if err == nil {
+		t.Fatal("expected empty-output error")
+	}
+	if !strings.Contains(err.Error(), "empty transcript") {
+		t.Fatalf("unexpected error: %v", err)
+	}
+
+	err = executeTrim(
+		"--input-file", input,
+		"--output-file", output,
+		"--remove", "1-4",
+		"--allow-empty",
+	)
+	if err != nil {
+		t.Fatalf("trim with --allow-empty failed: %v", err)
+	}
+
+	var transcript schema.Transcript
+	readJSON(t, output, &transcript)
+	if len(transcript.Segments) != 0 {
+		t.Fatalf("segment count = %d, want 0", len(transcript.Segments))
+	}
+}
+
+func TestTrimRejectsNonSeriatimInputArtifacts(t *testing.T) {
+	dir := t.TempDir()
+	input := writeJSONFile(t, dir, "raw-whisperx.json", `{
+		"segments": [
+			{"start": 1, "end": 2, "text": "hello"}
+		]
+	}`)
+	output := filepath.Join(dir, "trimmed.json")
+
+	err := executeTrim(
+		"--input-file", input,
+		"--output-file", output,
+		"--keep", "1",
+	)
+	if err == nil {
+		t.Fatal("expected invalid artifact error")
+	}
+	if !strings.Contains(err.Error(), "not a valid seriatim output artifact") {
+		t.Fatalf("unexpected error: %v", err)
+	}
+}
+
+func TestTrimReportFileContainsAuditFields(t *testing.T) {
+	dir := t.TempDir()
+	input := writeTrimFullFixture(t, dir, "input.json")
+	output := filepath.Join(dir, "trimmed.json")
+	reportPath := filepath.Join(dir, "trim-report.json")
+
+	err := executeTrim(
+		"--input-file", input,
+		"--output-file", output,
+		"--report-file", reportPath,
+		"--remove", "4,2",
+	)
+	if err != nil {
+		t.Fatalf("trim failed: %v", err)
+	}
+
+	var rpt report.Report
+	readJSON(t, reportPath, &rpt)
+	if len(rpt.Events) == 0 {
+		t.Fatal("expected report events")
+	}
+	if !hasReportEvent(rpt, "trim", "trim", "trimmed 4 input segment(s) into 2 output segment(s) with mode=remove") {
+		t.Fatal("expected trim summary event")
+	}
+	if !hasReportEvent(rpt, "trim", "validate-output", "validated 2 output segment(s)") {
+		t.Fatal("expected validation event")
+	}
+
+	audit := extractTrimAuditEvent(t, rpt)
+	if audit.Operation != "trim" {
+		t.Fatalf("operation = %q, want trim", audit.Operation)
+	}
+	if audit.InputFile != input {
+		t.Fatalf("input_file = %q, want %q", audit.InputFile, input)
+	}
+	if audit.OutputFile != output {
+		t.Fatalf("output_file = %q, want %q", audit.OutputFile, output)
+	}
+	if audit.InputSchema != config.OutputSchemaFull || audit.OutputSchema != config.OutputSchemaFull {
+		t.Fatalf("schemas = %q -> %q, want full -> full", audit.InputSchema, audit.OutputSchema)
+	}
+	if audit.Mode != "remove" {
+		t.Fatalf("mode = %q, want remove", audit.Mode)
+	}
+	if audit.Selector != "4,2" {
+		t.Fatalf("selector = %q, want %q", audit.Selector, "4,2")
+	}
+	assertIntSliceEqual(t, audit.SelectedIDs, []int{2, 4})
+	if audit.AllowEmpty {
+		t.Fatal("allow_empty should be false")
+	}
+	if audit.InputSegmentCount != 4 || audit.RetainedSegmentCount != 2 || audit.RemovedSegmentCount != 2 {
+		t.Fatalf("counts = input:%d retained:%d removed:%d, want 4/2/2", audit.InputSegmentCount, audit.RetainedSegmentCount, audit.RemovedSegmentCount)
+	}
+	assertIntSliceEqual(t, audit.RemovedInputIDs, []int{2, 4})
+	if len(audit.OldToNewIDMapping) != 2 {
+		t.Fatalf("mapping length = %d, want 2", len(audit.OldToNewIDMapping))
+	}
+	if audit.OldToNewIDMapping[0].OldID != 1 || audit.OldToNewIDMapping[0].NewID != 1 {
+		t.Fatalf("mapping[0] = %#v, want old_id=1 new_id=1", audit.OldToNewIDMapping[0])
+	}
+	if audit.OldToNewIDMapping[1].OldID != 3 || audit.OldToNewIDMapping[1].NewID != 2 {
+		t.Fatalf("mapping[1] = %#v, want old_id=3 new_id=2", audit.OldToNewIDMapping[1])
+	}
+	if !audit.OverlapGroupsRecomputed {
+		t.Fatal("expected overlap_groups_recomputed=true for full schema trim")
+	}
+}
+
+func TestTrimReportOldToNewMappingIsDeterministicSorted(t *testing.T) {
+	dir := t.TempDir()
+	input := writeTrimFullFixture(t, dir, "input.json")
+	output := filepath.Join(dir, "trimmed.json")
+	reportPath := filepath.Join(dir, "trim-report.json")
+
+	err := executeTrim(
+		"--input-file", input,
+		"--output-file", output,
+		"--report-file", reportPath,
+		"--keep", "4,1,3",
+	)
+	if err != nil {
+		t.Fatalf("trim failed: %v", err)
+	}
+
+	var rpt report.Report
+	readJSON(t, reportPath, &rpt)
+	audit := extractTrimAuditEvent(t, rpt)
+	if len(audit.OldToNewIDMapping) != 3 {
+		t.Fatalf("mapping length = %d, want 3", len(audit.OldToNewIDMapping))
+	}
+	for index, expectedOld := range []int{1, 3, 4} {
+		if audit.OldToNewIDMapping[index].OldID != expectedOld {
+			t.Fatalf("mapping[%d].old_id = %d, want %d", index, audit.OldToNewIDMapping[index].OldID, expectedOld)
+		}
+	}
+}
+
+func TestTrimNoReportFileWhenOmitted(t *testing.T) {
+	dir := t.TempDir()
+	input := writeTrimFullFixture(t, dir, "input.json")
+	output := filepath.Join(dir, "trimmed.json")
+	reportPath := filepath.Join(dir, "trim-report.json")
+
+	err := executeTrim(
+		"--input-file", input,
+		"--output-file", output,
+		"--keep", "1",
+	)
+	if err != nil {
+		t.Fatalf("trim failed: %v", err)
+	}
+
+	_, statErr := os.Stat(reportPath)
+	if !os.IsNotExist(statErr) {
+		t.Fatalf("expected no report file at %q, got err=%v", reportPath, statErr)
+	}
+}
+
+func executeTrim(args ...string) error {
+	cmd := NewRootCommand()
+	cmd.SetArgs(append([]string{"trim"}, args...))
+	return cmd.Execute()
+}
+
+func writeTrimFullFixture(t *testing.T, dir string, name string) string {
+	t.Helper()
+
+	first := 10
+	second := 20
+	third := 30
+	fourth := 40
+	value := schema.Transcript{
+		Metadata: schema.Metadata{
+			Application:           "seriatim",
+			Version:               "v-test",
+			InputReader:           "json-files",
+			InputFiles:            []string{"a.json"},
+			PreprocessingModules:  []string{"validate-raw"},
+			PostprocessingModules: []string{"assign-ids"},
+			OutputModules:         []string{"json"},
+		},
+		Segments: []schema.Segment{
+			{ID: 1, Source: "a.json", SourceSegmentIndex: &first, SourceRef: "a.json#10", Speaker: "A", Start: 1, End: 2, Text: "one", OverlapGroupID: 9},
+			{ID: 2, Source: "a.json", SourceSegmentIndex: &second, SourceRef: "a.json#20", Speaker: "B", Start: 2, End: 3, Text: "two", OverlapGroupID: 9},
+			{ID: 3, Source: "a.json", SourceSegmentIndex: &third, SourceRef: "a.json#30", Speaker: "C", Start: 4, End: 5, Text: "three", OverlapGroupID: 10},
+			{ID: 4, Source: "a.json", SourceSegmentIndex: &fourth, SourceRef: "a.json#40", Speaker: "D", Start: 5, End: 6, Text: "four", OverlapGroupID: 10},
+		},
+		OverlapGroups: []schema.OverlapGroup{
+			{ID: 9, Start: 1, End: 3, Segments: []string{"a.json#10", "a.json#20"}, Speakers: []string{"A", "B"}, Class: "unknown", Resolution: "unresolved"},
+		},
+	}
+
+	return writeTrimArtifactFile(t, dir, name, value)
+}
+
+func writeTrimMinimalFixture(t *testing.T, dir string, name string) string {
+	t.Helper()
+
+	value := schema.MinimalTranscript{
+		Metadata: schema.MinimalMetadata{
+			Application:  "seriatim",
+			Version:      "v-test",
+			OutputSchema: config.OutputSchemaMinimal,
+		},
+		Segments: []schema.MinimalSegment{
+			{ID: 1, Start: 1, End: 2, Speaker: "A", Text: "one"},
+			{ID: 2, Start: 2, End: 3, Speaker: "B", Text: "two"},
+		},
+	}
+
+	return writeTrimArtifactFile(t, dir, name, value)
+}
+
+func writeTrimIntermediateFixture(t *testing.T, dir string, name string) string {
+	t.Helper()
+
+	value := schema.IntermediateTranscript{
+		Metadata: schema.IntermediateMetadata{
+			Application:  "seriatim",
+			Version:      "v-test",
+			OutputSchema: config.OutputSchemaIntermediate,
+		},
+		Segments: []schema.IntermediateSegment{
+			{ID: 1, Start: 1, End: 2, Speaker: "A", Text: "one", Categories: []string{"word-run"}},
+			{ID: 2, Start: 2, End: 3, Speaker: "B", Text: "two", Categories: []string{"filler", "backchannel"}},
+		},
+	}
+
+	return writeTrimArtifactFile(t, dir, name, value)
+}
+
+func writeTrimMinimalWithIDsFixture(t *testing.T, dir string, name string, ids []int) string {
+	t.Helper()
+
+	if len(ids) < 2 {
+		t.Fatalf("need at least two IDs, got %d", len(ids))
+	}
+	value := schema.MinimalTranscript{
+		Metadata: schema.MinimalMetadata{
+			Application:  "seriatim",
+			Version:      "v-test",
+			OutputSchema: config.OutputSchemaMinimal,
+		},
+		Segments: []schema.MinimalSegment{
+			{ID: ids[0], Start: 1, End: 2, Speaker: "A", Text: "one"},
+			{ID: ids[1], Start: 2, End: 3, Speaker: "B", Text: "two"},
+		},
+	}
+
+	return writeTrimArtifactFile(t, dir, name, value)
+}
+
+func writeTrimFullOverlapFixture(t *testing.T, dir string, name string) string {
+	t.Helper()
+
+	first := 10
+	second := 20
+	third := 30
+	value := schema.Transcript{
+		Metadata: schema.Metadata{
+			Application:           "seriatim",
+			Version:               "v-test",
+			InputReader:           "json-files",
+			InputFiles:            []string{"a.json"},
+			PreprocessingModules:  []string{"validate-raw"},
+			PostprocessingModules: []string{"detect-overlaps", "assign-ids"},
+			OutputModules:         []string{"json"},
+		},
+		Segments: []schema.Segment{
+			{ID: 1, Source: "a.json", SourceSegmentIndex: &first, SourceRef: "a.json#10", Speaker: "A", Start: 1, End: 3, Text: "one", OverlapGroupID: 5},
+			{ID: 2, Source: "a.json", SourceSegmentIndex: &second, SourceRef: "a.json#20", Speaker: "B", Start: 2, End: 4, Text: "two", OverlapGroupID: 5},
+			{ID: 3, Source: "a.json", SourceSegmentIndex: &third, SourceRef: "a.json#30", Speaker: "C", Start: 6, End: 7, Text: "three", OverlapGroupID: 6},
+		},
+		OverlapGroups: []schema.OverlapGroup{
+			{ID: 99, Start: 0, End: 100, Segments: []string{"stale"}, Speakers: []string{"stale"}, Class: "unknown", Resolution: "unresolved"},
+		},
+	}
+
+	return writeTrimArtifactFile(t, dir, name, value)
+}
+
+func writeTrimArtifactFile(t *testing.T, dir string, name string, value any) string {
+	t.Helper()
+
+	data, err := json.MarshalIndent(value, "", "  ")
+	if err != nil {
+		t.Fatalf("marshal fixture: %v", err)
+	}
+	path := filepath.Join(dir, name)
+	if err := os.WriteFile(path, append(data, '\n'), 0o600); err != nil {
+		t.Fatalf("write fixture: %v", err)
+	}
+	return path
+}
+
+func assertSequentialIDs(t *testing.T, ids []int) {
+	t.Helper()
+	for index, id := range ids {
+		want := index + 1
+		if id != want {
+			t.Fatalf("id at index %d = %d, want %d", index, id, want)
+		}
+	}
+}
+
+func extractTrimAuditEvent(t *testing.T, rpt report.Report) trimAuditReport {
+	t.Helper()
+
+	for _, event := range rpt.Events {
+		if event.Stage == "trim" && event.Module == "trim-audit" {
+			var audit trimAuditReport
+			if err := json.Unmarshal([]byte(event.Message), &audit); err != nil {
+				t.Fatalf("decode trim audit event: %v", err)
+			}
+			return audit
+		}
+	}
+	t.Fatal("missing trim-audit event")
+	return trimAuditReport{}
+}
+
+func assertIntSliceEqual(t *testing.T, got []int, want []int) {
+	t.Helper()
+	if len(got) != len(want) {
+		t.Fatalf("slice length = %d, want %d", len(got), len(want))
+	}
+	for index := range got {
+		if got[index] != want[index] {
+			t.Fatalf("slice[%d] = %d, want %d (full got=%v, want=%v)", index, got[index], want[index], got, want)
+		}
+	}
+}

internal/config/config.go

@@ -13,22 +13,23 @@ import (
 const (
 	DefaultInputReader            = "json-files"
 	DefaultOutputModules          = "json"
-	DefaultOutputSchema           = OutputSchemaDefault
+	DefaultOutputSchema           = OutputSchemaIntermediate
 	DefaultPreprocessingModules   = "validate-raw,normalize-speakers,trim-text"
 	DefaultPostprocessingModules  = "detect-overlaps,resolve-overlaps,backchannel,filler,resolve-danglers,coalesce,detect-overlaps,autocorrect,assign-ids,validate-output"
-	DefaultOverlapWordRunGap      = 0.75
+	DefaultOverlapWordRunGap      = 1.0
 	DefaultWordRunReorderWindow   = 1.0
 	DefaultCoalesceGap            = 3.0
 	DefaultCoalesceGapValue       = "3.0"
 	DefaultBackchannelMaxDuration = 2.0
 	DefaultFillerMaxDuration      = 1.25
+	OutputSchemaEnv               = "SERIATIM_OUTPUT_SCHEMA"
 	OverlapWordRunGapEnv          = "SERIATIM_OVERLAP_WORD_RUN_GAP"
 	WordRunReorderWindowEnv       = "SERIATIM_OVERLAP_WORD_RUN_REORDER_WINDOW"
 	BackchannelMaxDurationEnv     = "SERIATIM_BACKCHANNEL_MAX_DURATION"
 	FillerMaxDurationEnv          = "SERIATIM_FILLER_MAX_DURATION"
-	OutputSchemaDefault           = "default"
-	OutputSchemaSeriatim          = "seriatim"
-	OutputSchemaMinimal           = "minimal"
+	OutputSchemaMinimal           = "seriatim-minimal"
+	OutputSchemaIntermediate      = "seriatim-intermediate"
+	OutputSchemaFull              = "seriatim-full"
 )
 
 // MergeOptions captures raw CLI option values before validation.
@@ -46,6 +47,26 @@ type MergeOptions struct {
 	CoalesceGap           string
 }
 
+// TrimOptions captures raw CLI option values before validation.
+type TrimOptions struct {
+	InputFile    string
+	OutputFile   string
+	ReportFile   string
+	Keep         string
+	Remove       string
+	OutputSchema string
+	AllowEmpty   bool
+}
+
+// NormalizeOptions captures raw CLI option values before validation.
+type NormalizeOptions struct {
+	InputFile     string
+	OutputFile    string
+	ReportFile    string
+	OutputSchema  string
+	OutputModules string
+}
+
 // Config is the validated runtime configuration for a merge invocation.
 type Config struct {
 	InputFiles             []string
@@ -65,12 +86,32 @@ type Config struct {
 	FillerMaxDuration      float64
 }
 
+// TrimConfig is the validated runtime configuration for a trim invocation.
+type TrimConfig struct {
+	InputFile    string
+	OutputFile   string
+	ReportFile   string
+	Mode         string
+	Selector     string
+	OutputSchema string
+	AllowEmpty   bool
+}
+
+// NormalizeConfig is the validated runtime configuration for a normalize invocation.
+type NormalizeConfig struct {
+	InputFile     string
+	OutputFile    string
+	ReportFile    string
+	OutputSchema  string
+	OutputModules []string
+}
+
 // NewMergeConfig validates raw merge options and returns normalized config.
 func NewMergeConfig(opts MergeOptions) (Config, error) {
 	cfg := Config{
 		InputReader:            strings.TrimSpace(opts.InputReader),
 		OutputModules:          nil,
-		OutputSchema:           strings.TrimSpace(opts.OutputSchema),
+		OutputSchema:           "",
 		PreprocessingModules:   nil,
 		PostprocessingModules:  nil,
 		OverlapWordRunGap:      DefaultOverlapWordRunGap,
@@ -83,14 +124,12 @@ func NewMergeConfig(opts MergeOptions) (Config, error) {
 	if cfg.InputReader == "" {
 		return Config{}, errors.New("--input-reader is required")
 	}
-	if cfg.OutputSchema == "" {
-		cfg.OutputSchema = DefaultOutputSchema
-	}
-	if err := validateOutputSchema(cfg.OutputSchema); err != nil {
+	var err error
+	cfg.OutputSchema, err = resolveOutputSchema(opts.OutputSchema)
+	if err != nil {
 		return Config{}, err
 	}
 
-	var err error
 	cfg.OutputModules, err = parseModuleList(opts.OutputModules)
 	if err != nil {
 		return Config{}, fmt.Errorf("--output-modules: %w", err)
@@ -169,6 +208,111 @@ func NewMergeConfig(opts MergeOptions) (Config, error) {
 	return cfg, nil
 }
 
+// NewTrimConfig validates raw trim options and returns normalized config.
+func NewTrimConfig(opts TrimOptions) (TrimConfig, error) {
+	inputFile := filepath.Clean(strings.TrimSpace(opts.InputFile))
+	if strings.TrimSpace(opts.InputFile) == "" {
+		return TrimConfig{}, errors.New("--input-file is required")
+	}
+	if err := requireFile(inputFile, "--input-file"); err != nil {
+		return TrimConfig{}, err
+	}
+
+	outputFile, err := normalizeOutputPath(opts.OutputFile, "--output-file")
+	if err != nil {
+		return TrimConfig{}, err
+	}
+
+	reportFile := ""
+	if strings.TrimSpace(opts.ReportFile) != "" {
+		reportFile, err = normalizeOutputPath(opts.ReportFile, "--report-file")
+		if err != nil {
+			return TrimConfig{}, err
+		}
+	}
+
+	keep := strings.TrimSpace(opts.Keep)
+	remove := strings.TrimSpace(opts.Remove)
+	if keep == "" && remove == "" {
+		return TrimConfig{}, errors.New("exactly one of --keep or --remove is required")
+	}
+	if keep != "" && remove != "" {
+		return TrimConfig{}, errors.New("--keep and --remove are mutually exclusive")
+	}
+
+	mode := "keep"
+	selector := keep
+	if remove != "" {
+		mode = "remove"
+		selector = remove
+	}
+
+	outputSchema := strings.TrimSpace(opts.OutputSchema)
+	if outputSchema != "" {
+		if err := validateOutputSchema(outputSchema); err != nil {
+			return TrimConfig{}, err
+		}
+	}
+
+	return TrimConfig{
+		InputFile:    inputFile,
+		OutputFile:   outputFile,
+		ReportFile:   reportFile,
+		Mode:         mode,
+		Selector:     selector,
+		OutputSchema: outputSchema,
+		AllowEmpty:   opts.AllowEmpty,
+	}, nil
+}
+
+// NewNormalizeConfig validates raw normalize options and returns normalized config.
+func NewNormalizeConfig(opts NormalizeOptions) (NormalizeConfig, error) {
+	inputFile := filepath.Clean(strings.TrimSpace(opts.InputFile))
+	if strings.TrimSpace(opts.InputFile) == "" {
+		return NormalizeConfig{}, errors.New("--input-file is required")
+	}
+	if err := requireFile(inputFile, "--input-file"); err != nil {
+		return NormalizeConfig{}, err
+	}
+
+	outputFile, err := normalizeOutputPath(opts.OutputFile, "--output-file")
+	if err != nil {
+		return NormalizeConfig{}, err
+	}
+
+	reportFile := ""
+	if strings.TrimSpace(opts.ReportFile) != "" {
+		reportFile, err = normalizeOutputPath(opts.ReportFile, "--report-file")
+		if err != nil {
+			return NormalizeConfig{}, err
+		}
+	}
+
+	outputSchema, err := resolveOutputSchema(opts.OutputSchema)
+	if err != nil {
+		return NormalizeConfig{}, err
+	}
+
+	outputModules, err := parseModuleList(opts.OutputModules)
+	if err != nil {
+		return NormalizeConfig{}, fmt.Errorf("--output-modules: %w", err)
+	}
+	if len(outputModules) == 0 {
+		return NormalizeConfig{}, errors.New("--output-modules must include at least one module")
+	}
+	if err := validateNormalizeOutputModules(outputModules); err != nil {
+		return NormalizeConfig{}, err
+	}
+
+	return NormalizeConfig{
+		InputFile:     inputFile,
+		OutputFile:    outputFile,
+		ReportFile:    reportFile,
+		OutputSchema:  outputSchema,
+		OutputModules: outputModules,
+	}, nil
+}
+
 func parseModuleList(value string) ([]string, error) {
 	value = strings.TrimSpace(value)
 	if value == "" {
@@ -189,13 +333,27 @@ func parseModuleList(value string) ([]string, error) {
 
 func validateOutputSchema(value string) error {
 	switch value {
-	case OutputSchemaDefault, OutputSchemaSeriatim, OutputSchemaMinimal:
+	case OutputSchemaMinimal, OutputSchemaIntermediate, OutputSchemaFull:
 		return nil
 	default:
-		return fmt.Errorf("--output-schema must be one of %q, %q, or %q", OutputSchemaDefault, OutputSchemaMinimal, OutputSchemaSeriatim)
+		return fmt.Errorf("--output-schema must be one of %q, %q, or %q", OutputSchemaMinimal, OutputSchemaIntermediate, OutputSchemaFull)
 	}
 }
 
+func resolveOutputSchema(value string) (string, error) {
+	value = strings.TrimSpace(value)
+	if value == "" {
+		value = strings.TrimSpace(os.Getenv(OutputSchemaEnv))
+	}
+	if value == "" {
+		value = DefaultOutputSchema
+	}
+	if err := validateOutputSchema(value); err != nil {
+		return "", err
+	}
+	return value, nil
+}
+
 func normalizeInputFiles(paths []string) ([]string, error) {
 	if len(paths) == 0 {
 		return nil, errors.New("at least one --input-file is required")
@@ -308,3 +466,12 @@ func contains(values []string, target string) bool {
 	}
 	return false
 }
+
+func validateNormalizeOutputModules(modules []string) error {
+	for _, module := range modules {
+		if module != "json" {
+			return fmt.Errorf("unknown output module %q", module)
+		}
+	}
+	return nil
+}

internal/config/config_test.go

@@ -46,7 +46,8 @@ func TestDuplicateInputFilesFailValidation(t *testing.T) {
 	}
 }
 
-func TestOutputSchemaDefaultsToDefault(t *testing.T) {
+func TestOutputSchemaDefaultsToIntermediate(t *testing.T) {
+	t.Setenv(OutputSchemaEnv, "")
 	dir := t.TempDir()
 	input := writeTempFile(t, dir, "input.json")
 	output := filepath.Join(dir, "merged.json")
@@ -67,7 +68,8 @@ func TestOutputSchemaDefaultsToDefault(t *testing.T) {
 	}
 }
 
-func TestOutputSchemaAcceptsDefault(t *testing.T) {
+func TestOutputSchemaAcceptsIntermediate(t *testing.T) {
+	t.Setenv(OutputSchemaEnv, "")
 	dir := t.TempDir()
 	input := writeTempFile(t, dir, "input.json")
 	output := filepath.Join(dir, "merged.json")
@@ -77,19 +79,20 @@ func TestOutputSchemaAcceptsDefault(t *testing.T) {
 		OutputFile:            output,
 		InputReader:           DefaultInputReader,
 		OutputModules:         DefaultOutputModules,
-		OutputSchema:          OutputSchemaDefault,
+		OutputSchema:          OutputSchemaIntermediate,
 		PreprocessingModules:  DefaultPreprocessingModules,
 		PostprocessingModules: DefaultPostprocessingModules,
 	})
 	if err != nil {
 		t.Fatalf("config failed: %v", err)
 	}
-	if cfg.OutputSchema != OutputSchemaDefault {
-		t.Fatalf("output schema = %q, want %q", cfg.OutputSchema, OutputSchemaDefault)
+	if cfg.OutputSchema != OutputSchemaIntermediate {
+		t.Fatalf("output schema = %q, want %q", cfg.OutputSchema, OutputSchemaIntermediate)
 	}
 }
 
 func TestOutputSchemaAcceptsMinimal(t *testing.T) {
+	t.Setenv(OutputSchemaEnv, "")
 	dir := t.TempDir()
 	input := writeTempFile(t, dir, "input.json")
 	output := filepath.Join(dir, "merged.json")
@@ -111,7 +114,8 @@ func TestOutputSchemaAcceptsMinimal(t *testing.T) {
 	}
 }
 
-func TestOutputSchemaAcceptsSeriatim(t *testing.T) {
+func TestOutputSchemaAcceptsFull(t *testing.T) {
+	t.Setenv(OutputSchemaEnv, "")
 	dir := t.TempDir()
 	input := writeTempFile(t, dir, "input.json")
 	output := filepath.Join(dir, "merged.json")
@@ -121,19 +125,87 @@ func TestOutputSchemaAcceptsSeriatim(t *testing.T) {
 		OutputFile:            output,
 		InputReader:           DefaultInputReader,
 		OutputModules:         DefaultOutputModules,
-		OutputSchema:          OutputSchemaSeriatim,
+		OutputSchema:          OutputSchemaFull,
 		PreprocessingModules:  DefaultPreprocessingModules,
 		PostprocessingModules: DefaultPostprocessingModules,
 	})
 	if err != nil {
 		t.Fatalf("config failed: %v", err)
 	}
-	if cfg.OutputSchema != OutputSchemaSeriatim {
-		t.Fatalf("output schema = %q, want %q", cfg.OutputSchema, OutputSchemaSeriatim)
+	if cfg.OutputSchema != OutputSchemaFull {
+		t.Fatalf("output schema = %q, want %q", cfg.OutputSchema, OutputSchemaFull)
+	}
+}
+
+func TestOutputSchemaUsesEnvWhenFlagOmitted(t *testing.T) {
+	t.Setenv(OutputSchemaEnv, OutputSchemaFull)
+	dir := t.TempDir()
+	input := writeTempFile(t, dir, "input.json")
+	output := filepath.Join(dir, "merged.json")
+
+	cfg, err := NewMergeConfig(MergeOptions{
+		InputFiles:            []string{input},
+		OutputFile:            output,
+		InputReader:           DefaultInputReader,
+		OutputModules:         DefaultOutputModules,
+		PreprocessingModules:  DefaultPreprocessingModules,
+		PostprocessingModules: DefaultPostprocessingModules,
+	})
+	if err != nil {
+		t.Fatalf("config failed: %v", err)
+	}
+	if cfg.OutputSchema != OutputSchemaFull {
+		t.Fatalf("output schema = %q, want %q", cfg.OutputSchema, OutputSchemaFull)
+	}
+}
+
+func TestOutputSchemaFlagOverridesEnv(t *testing.T) {
+	t.Setenv(OutputSchemaEnv, OutputSchemaFull)
+	dir := t.TempDir()
+	input := writeTempFile(t, dir, "input.json")
+	output := filepath.Join(dir, "merged.json")
+
+	cfg, err := NewMergeConfig(MergeOptions{
+		InputFiles:            []string{input},
+		OutputFile:            output,
+		InputReader:           DefaultInputReader,
+		OutputModules:         DefaultOutputModules,
+		OutputSchema:          OutputSchemaMinimal,
+		PreprocessingModules:  DefaultPreprocessingModules,
+		PostprocessingModules: DefaultPostprocessingModules,
+	})
+	if err != nil {
+		t.Fatalf("config failed: %v", err)
+	}
+	if cfg.OutputSchema != OutputSchemaMinimal {
+		t.Fatalf("output schema = %q, want %q", cfg.OutputSchema, OutputSchemaMinimal)
+	}
+}
+
+func TestOutputSchemaRejectsInvalidEnvValue(t *testing.T) {
+	t.Setenv(OutputSchemaEnv, "compact")
+	dir := t.TempDir()
+	input := writeTempFile(t, dir, "input.json")
+	output := filepath.Join(dir, "merged.json")
+
+	_, err := NewMergeConfig(MergeOptions{
+		InputFiles:            []string{input},
+		OutputFile:            output,
+		InputReader:           DefaultInputReader,
+		OutputModules:         DefaultOutputModules,
+		PreprocessingModules:  DefaultPreprocessingModules,
+		PostprocessingModules: DefaultPostprocessingModules,
+	})
+	if err == nil {
+		t.Fatal("expected output schema error")
+	}
+	if !strings.Contains(err.Error(), "--output-schema must be one of") {
+		t.Fatalf("unexpected error: %v", err)
 	}
 }
 
 func TestOutputSchemaRejectsUnknownValue(t *testing.T) {
+	t.Setenv(OutputSchemaEnv, "")
 	dir := t.TempDir()
 	input := writeTempFile(t, dir, "input.json")
 	output := filepath.Join(dir, "merged.json")
@@ -155,7 +227,36 @@ func TestOutputSchemaRejectsUnknownValue(t *testing.T) {
 	}
 }
 
-func TestOverlapWordRunGapDefaultsTo075(t *testing.T) {
+func TestOutputSchemaRejectsLegacyValues(t *testing.T) {
+	tests := []string{"default", "minimal", "seriatim"}
+
+	for _, legacy := range tests {
+		t.Run(legacy, func(t *testing.T) {
+			t.Setenv(OutputSchemaEnv, "")
+			dir := t.TempDir()
+			input := writeTempFile(t, dir, "input.json")
+			output := filepath.Join(dir, "merged.json")
+
+			_, err := NewMergeConfig(MergeOptions{
+				InputFiles:            []string{input},
+				OutputFile:            output,
+				InputReader:           DefaultInputReader,
+				OutputModules:         DefaultOutputModules,
+				OutputSchema:          legacy,
+				PreprocessingModules:  DefaultPreprocessingModules,
+				PostprocessingModules: DefaultPostprocessingModules,
+			})
+			if err == nil {
+				t.Fatal("expected output schema error")
+			}
+			if !strings.Contains(err.Error(), "--output-schema must be one of") {
+				t.Fatalf("unexpected error: %v", err)
+			}
+		})
+	}
+}
+
+func TestOverlapWordRunGapDefaultsTo1(t *testing.T) {
 	t.Setenv(OverlapWordRunGapEnv, "")
 	dir := t.TempDir()
 	input := writeTempFile(t, dir, "input.json")
@@ -511,6 +612,206 @@ func TestCoalesceGapRejectsInvalidOverride(t *testing.T) {
 	}
 }
 
+func TestNewTrimConfigRequiresInputAndOutput(t *testing.T) {
+	dir := t.TempDir()
+	input := writeTempFile(t, dir, "input.json")
+	output := filepath.Join(dir, "trimmed.json")
+
+	_, err := NewTrimConfig(TrimOptions{
+		OutputFile: output,
+		Keep:       "1",
+	})
+	if err == nil || !strings.Contains(err.Error(), "--input-file is required") {
+		t.Fatalf("expected input-file required error, got %v", err)
+	}
+
+	_, err = NewTrimConfig(TrimOptions{
+		InputFile: input,
+		Keep:      "1",
+	})
+	if err == nil || !strings.Contains(err.Error(), "--output-file is required") {
+		t.Fatalf("expected output-file required error, got %v", err)
+	}
+}
+
+func TestNewTrimConfigRequiresExactlyOneSelectorFlag(t *testing.T) {
+	dir := t.TempDir()
+	input := writeTempFile(t, dir, "input.json")
+	output := filepath.Join(dir, "trimmed.json")
+
+	_, err := NewTrimConfig(TrimOptions{
+		InputFile:  input,
+		OutputFile: output,
+	})
+	if err == nil || !strings.Contains(err.Error(), "exactly one of --keep or --remove is required") {
+		t.Fatalf("expected missing selector error, got %v", err)
+	}
+
+	_, err = NewTrimConfig(TrimOptions{
+		InputFile:  input,
+		OutputFile: output,
+		Keep:       "1",
+		Remove:     "2",
+	})
+	if err == nil || !strings.Contains(err.Error(), "mutually exclusive") {
+		t.Fatalf("expected mutually exclusive selector error, got %v", err)
+	}
+}
+
+func TestNewTrimConfigAcceptsOutputSchemaOverride(t *testing.T) {
+	dir := t.TempDir()
+	input := writeTempFile(t, dir, "input.json")
+	output := filepath.Join(dir, "trimmed.json")
+	reportPath := filepath.Join(dir, "report.json")
+
+	cfg, err := NewTrimConfig(TrimOptions{
+		InputFile:    input,
+		OutputFile:   output,
+		ReportFile:   reportPath,
+		Remove:       "3-5",
+		OutputSchema: OutputSchemaMinimal,
+		AllowEmpty:   true,
+	})
+	if err != nil {
+		t.Fatalf("config failed: %v", err)
+	}
+	if cfg.Mode != "remove" {
+		t.Fatalf("mode = %q, want remove", cfg.Mode)
+	}
+	if cfg.Selector != "3-5" {
+		t.Fatalf("selector = %q, want 3-5", cfg.Selector)
+	}
+	if cfg.OutputSchema != OutputSchemaMinimal {
+		t.Fatalf("output schema = %q, want %q", cfg.OutputSchema, OutputSchemaMinimal)
+	}
+	if !cfg.AllowEmpty {
+		t.Fatal("allow empty should be true")
+	}
+	if cfg.ReportFile != reportPath {
+		t.Fatalf("report file = %q, want %q", cfg.ReportFile, reportPath)
+	}
+}
+
+func TestNewTrimConfigRejectsInvalidOutputSchemaOverride(t *testing.T) {
+	dir := t.TempDir()
+	input := writeTempFile(t, dir, "input.json")
+	output := filepath.Join(dir, "trimmed.json")
+
+	_, err := NewTrimConfig(TrimOptions{
+		InputFile:    input,
+		OutputFile:   output,
+		Keep:         "1",
+		OutputSchema: "compact",
+	})
+	if err == nil {
+		t.Fatal("expected output schema validation error")
+	}
+	if !strings.Contains(err.Error(), "--output-schema must be one of") {
+		t.Fatalf("unexpected error: %v", err)
+	}
+}
+
+func TestNewNormalizeConfigRequiresInputFile(t *testing.T) {
+	dir := t.TempDir()
+	output := filepath.Join(dir, "normalized.json")
+
+	_, err := NewNormalizeConfig(NormalizeOptions{
+		OutputFile:    output,
+		OutputModules: DefaultOutputModules,
+	})
+	if err == nil {
+		t.Fatal("expected input-file required error")
+	}
+	if !strings.Contains(err.Error(), "--input-file is required") {
+		t.Fatalf("unexpected error: %v", err)
+	}
+}
+
+func TestNewNormalizeConfigRequiresOutputFile(t *testing.T) {
+	dir := t.TempDir()
+	input := writeTempFile(t, dir, "input.json")
+
+	_, err := NewNormalizeConfig(NormalizeOptions{
+		InputFile:     input,
+		OutputModules: DefaultOutputModules,
+	})
+	if err == nil {
+		t.Fatal("expected output-file required error")
+	}
+	if !strings.Contains(err.Error(), "--output-file is required") {
+		t.Fatalf("unexpected error: %v", err)
+	}
+}
+
+func TestNewNormalizeConfigResolvesOutputSchemaDefaultAndEnv(t *testing.T) {
+	dir := t.TempDir()
+	input := writeTempFile(t, dir, "input.json")
+	output := filepath.Join(dir, "normalized.json")
+
+	t.Setenv(OutputSchemaEnv, "")
+	cfg, err := NewNormalizeConfig(NormalizeOptions{
+		InputFile:     input,
+		OutputFile:    output,
+		OutputModules: DefaultOutputModules,
+	})
+	if err != nil {
+		t.Fatalf("config failed: %v", err)
+	}
+	if cfg.OutputSchema != DefaultOutputSchema {
+		t.Fatalf("output schema = %q, want %q", cfg.OutputSchema, DefaultOutputSchema)
+	}
+
+	t.Setenv(OutputSchemaEnv, OutputSchemaMinimal)
+	cfg, err = NewNormalizeConfig(NormalizeOptions{
+		InputFile:     input,
+		OutputFile:    output,
+		OutputModules: DefaultOutputModules,
+	})
+	if err != nil {
+		t.Fatalf("config failed: %v", err)
+	}
+	if cfg.OutputSchema != OutputSchemaMinimal {
+		t.Fatalf("output schema = %q, want %q", cfg.OutputSchema, OutputSchemaMinimal)
+	}
+}
+
+func TestNewNormalizeConfigRejectsInvalidOutputSchema(t *testing.T) {
+	dir := t.TempDir()
+	input := writeTempFile(t, dir, "input.json")
+	output := filepath.Join(dir, "normalized.json")
+
+	_, err := NewNormalizeConfig(NormalizeOptions{
+		InputFile:     input,
+		OutputFile:    output,
+		OutputSchema:  "compact",
+		OutputModules: DefaultOutputModules,
+	})
+	if err == nil {
+		t.Fatal("expected output schema error")
+	}
+	if !strings.Contains(err.Error(), "--output-schema must be one of") {
+		t.Fatalf("unexpected error: %v", err)
+	}
+}
+
+func TestNewNormalizeConfigRejectsUnknownOutputModule(t *testing.T) {
+	dir := t.TempDir()
+	input := writeTempFile(t, dir, "input.json")
+	output := filepath.Join(dir, "normalized.json")
+
+	_, err := NewNormalizeConfig(NormalizeOptions{
+		InputFile:     input,
+		OutputFile:    output,
+		OutputModules: "json,yaml",
+	})
+	if err == nil {
+		t.Fatal("expected output module error")
+	}
+	if !strings.Contains(err.Error(), "unknown output module") {
+		t.Fatalf("unexpected error: %v", err)
+	}
+}
+
 func assertPositiveFloatEnvValidation(t *testing.T, envName string) {
 	t.Helper()
 

internal/danglers/danglers.go

@@ -184,7 +184,7 @@ func nearestPriorMatch(segments []model.Segment, consumed []bool, index int) int
 		if consumed[candidate] {
 			continue
 		}
-		if sharesDerivedFrom(segments[index], segments[candidate]) {
+		if segments[index].Speaker == segments[candidate].Speaker && sharesDerivedFrom(segments[index], segments[candidate]) {
 			return candidate
 		}
 	}
@@ -196,7 +196,7 @@ func nearestSubsequentMatch(segments []model.Segment, consumed []bool, index int
 		if consumed[candidate] {
 			continue
 		}
-		if sharesDerivedFrom(segments[index], segments[candidate]) {
+		if segments[index].Speaker == segments[candidate].Speaker && sharesDerivedFrom(segments[index], segments[candidate]) {
 			return candidate
 		}
 	}

internal/danglers/danglers_test.go

@@ -51,6 +51,30 @@ func TestApplyUsesAnyDerivedFromIntersection(t *testing.T) {
 	assertSegment(t, got.Segments[0], "resolve-danglers:1", "target end.", 1, 4, []string{"source#1", "source#2", "source#3"})
 }
 
+func TestApplyDoesNotMergeDanglersAcrossSpeakersBackward(t *testing.T) {
+	in := transcript(
+		segment("a", "Alice", 1, 2, "target", []string{"source#1"}),
+		segment("b", "Bob", 3, 4, "end.", []string{"source#1"}),
+	)
+
+	got, summary := Apply(in)
+	if summary.DanglersMerged != 0 || !reflect.DeepEqual(got, in) {
+		t.Fatalf("unexpected merge:\ngot  %#v\nwant %#v", got, in)
+	}
+}
+
+func TestApplyDoesNotMergeDanglersAcrossSpeakersForward(t *testing.T) {
+	in := transcript(
+		segment("a", "Alice", 1, 2, "start", []string{"source#1"}),
+		segment("b", "Bob", 3, 4, "target", []string{"source#1"}),
+	)
+
+	got, summary := Apply(in)
+	if summary.DanglersMerged != 0 || !reflect.DeepEqual(got, in) {
+		t.Fatalf("unexpected merge:\ngot  %#v\nwant %#v", got, in)
+	}
+}
+
 func TestApplyDoesNotMergeWithoutSharedProvenance(t *testing.T) {
 	in := transcript(
 		segment("a", "Alice", 1, 2, "target", []string{"source#1"}),

internal/normalize/build.go

@@ -0,0 +1,216 @@
+package normalize
+
+import (
+	"fmt"
+	"path/filepath"
+	"sort"
+	"strings"
+
+	"gitea.maximumdirect.net/eric/seriatim/internal/artifact"
+	"gitea.maximumdirect.net/eric/seriatim/internal/buildinfo"
+	"gitea.maximumdirect.net/eric/seriatim/internal/config"
+	"gitea.maximumdirect.net/eric/seriatim/schema"
+)
+
+// BuildResult contains normalize output plus deterministic transformation diagnostics.
+type BuildResult struct {
+	Output                 any
+	SortingChanged         bool
+	IDsReassigned          bool
+	SegmentsWithCategories int
+}
+
+// Build converts parsed normalize input into a selected seriatim output schema.
+func Build(parsed ParsedTranscript, cfg config.NormalizeConfig) (BuildResult, error) {
+	ordered := sortedSegments(parsed.Segments)
+	sortingChanged := didSortingChangeOrder(ordered)
+	idsReassigned := didReassignIDs(ordered)
+	segmentsWithCategories := countSegmentsWithCategories(ordered)
+
+	switch cfg.OutputSchema {
+	case config.OutputSchemaMinimal:
+		output := buildMinimal(ordered)
+		if err := schema.ValidateMinimalTranscript(output); err != nil {
+			return BuildResult{}, fmt.Errorf("validate normalize output: %w", err)
+		}
+		return BuildResult{
+			Output:                 output,
+			SortingChanged:         sortingChanged,
+			IDsReassigned:          idsReassigned,
+			SegmentsWithCategories: segmentsWithCategories,
+		}, nil
+	case config.OutputSchemaIntermediate:
+		output := buildIntermediate(ordered)
+		if err := schema.ValidateIntermediateTranscript(output); err != nil {
+			return BuildResult{}, fmt.Errorf("validate normalize output: %w", err)
+		}
+		return BuildResult{
+			Output:                 output,
+			SortingChanged:         sortingChanged,
+			IDsReassigned:          idsReassigned,
+			SegmentsWithCategories: segmentsWithCategories,
+		}, nil
+	case config.OutputSchemaFull:
+		output := buildFull(ordered, cfg)
+		if err := schema.ValidateTranscript(output); err != nil {
+			return BuildResult{}, fmt.Errorf("validate normalize output: %w", err)
+		}
+		return BuildResult{
+			Output:                 output,
+			SortingChanged:         sortingChanged,
+			IDsReassigned:          idsReassigned,
+			SegmentsWithCategories: segmentsWithCategories,
+		}, nil
+	default:
+		return BuildResult{}, fmt.Errorf("unsupported output schema %q", cfg.OutputSchema)
+	}
+}
+
+func sortedSegments(input []InputSegment) []InputSegment {
+	ordered := make([]InputSegment, len(input))
+	copy(ordered, input)
+	sort.SliceStable(ordered, func(i, j int) bool {
+		left := ordered[i]
+		right := ordered[j]
+		if left.Start != right.Start {
+			return left.Start < right.Start
+		}
+		if left.End != right.End {
+			return left.End < right.End
+		}
+		if left.InputIndex != right.InputIndex {
+			return left.InputIndex < right.InputIndex
+		}
+		return left.Speaker < right.Speaker
+	})
+	return ordered
+}
+
+func buildMinimal(segments []InputSegment) schema.MinimalTranscript {
+	outputSegments := make([]schema.MinimalSegment, len(segments))
+	for index, segment := range segments {
+		outputSegments[index] = schema.MinimalSegment{
+			ID:      index + 1,
+			Start:   segment.Start,
+			End:     segment.End,
+			Speaker: segment.Speaker,
+			Text:    segment.Text,
+		}
+	}
+
+	return schema.MinimalTranscript{
+		Metadata: schema.MinimalMetadata{
+			Application:  artifact.ApplicationName,
+			Version:      buildinfo.Version,
+			OutputSchema: config.OutputSchemaMinimal,
+		},
+		Segments: outputSegments,
+	}
+}
+
+func buildIntermediate(segments []InputSegment) schema.IntermediateTranscript {
+	outputSegments := make([]schema.IntermediateSegment, len(segments))
+	for index, segment := range segments {
+		outputSegments[index] = schema.IntermediateSegment{
+			ID:         index + 1,
+			Start:      segment.Start,
+			End:        segment.End,
+			Speaker:    segment.Speaker,
+			Text:       segment.Text,
+			Categories: append([]string(nil), segment.Categories...),
+		}
+	}
+
+	return schema.IntermediateTranscript{
+		Metadata: schema.IntermediateMetadata{
+			Application:  artifact.ApplicationName,
+			Version:      buildinfo.Version,
+			OutputSchema: config.OutputSchemaIntermediate,
+		},
+		Segments: outputSegments,
+	}
+}
+
+func buildFull(segments []InputSegment, cfg config.NormalizeConfig) schema.Transcript {
+	defaultSource := filepath.Base(cfg.InputFile)
+	outputSegments := make([]schema.Segment, len(segments))
+	for index, segment := range segments {
+		source := strings.TrimSpace(segment.Source)
+		if source == "" {
+			source = defaultSource
+		}
+
+		sourceSegmentIndex := copyIntPtr(segment.SourceSegmentIndex)
+		if sourceSegmentIndex == nil {
+			fallback := segment.InputIndex
+			sourceSegmentIndex = &fallback
+		}
+
+		outputSegments[index] = schema.Segment{
+			ID:                 index + 1,
+			Source:             source,
+			SourceSegmentIndex: sourceSegmentIndex,
+			SourceRef:          segment.SourceRef,
+			DerivedFrom:        append([]string(nil), segment.DerivedFrom...),
+			Speaker:            segment.Speaker,
+			Start:              segment.Start,
+			End:                segment.End,
+			Text:               segment.Text,
+			Categories:         append([]string(nil), segment.Categories...),
+		}
+	}
+
+	return schema.Transcript{
+		Metadata: schema.Metadata{
+			Application:           artifact.ApplicationName,
+			Version:               buildinfo.Version,
+			InputReader:           "normalize-input",
+			InputFiles:            []string{cfg.InputFile},
+			PreprocessingModules:  []string{},
+			PostprocessingModules: []string{},
+			OutputModules:         append([]string(nil), cfg.OutputModules...),
+		},
+		Segments:      outputSegments,
+		OverlapGroups: []schema.OverlapGroup{},
+	}
+}
+
+func copyIntPtr(value *int) *int {
+	if value == nil {
+		return nil
+	}
+	copied := *value
+	return &copied
+}
+
+func didSortingChangeOrder(segments []InputSegment) bool {
+	for index, segment := range segments {
+		if segment.InputIndex != index {
+			return true
+		}
+	}
+	return false
+}
+
+func didReassignIDs(segments []InputSegment) bool {
+	if len(segments) == 0 {
+		return false
+	}
+	for index, segment := range segments {
+		newID := index + 1
+		if segment.OriginalID == nil || *segment.OriginalID != newID {
+			return true
+		}
+	}
+	return false
+}
+
+func countSegmentsWithCategories(segments []InputSegment) int {
+	count := 0
+	for _, segment := range segments {
+		if len(segment.Categories) > 0 {
+			count++
+		}
+	}
+	return count
+}

internal/normalize/normalize.go

@@ -0,0 +1,121 @@
+package normalize
+
+import (
+	"context"
+	"encoding/json"
+	"fmt"
+	"os"
+	"strings"
+
+	"gitea.maximumdirect.net/eric/seriatim/internal/artifact"
+	"gitea.maximumdirect.net/eric/seriatim/internal/buildinfo"
+	"gitea.maximumdirect.net/eric/seriatim/internal/config"
+	"gitea.maximumdirect.net/eric/seriatim/internal/report"
+)
+
+type normalizeAudit struct {
+	Command                string   `json:"command"`
+	InputFile              string   `json:"input_file"`
+	OutputFile             string   `json:"output_file"`
+	InputShape             string   `json:"input_shape"`
+	InputSegmentCount      int      `json:"input_segment_count"`
+	OutputSchema           string   `json:"output_schema"`
+	OutputModules          []string `json:"output_modules"`
+	IDsReassigned          bool     `json:"ids_reassigned"`
+	SortingChangedInput    bool     `json:"sorting_changed_input_order"`
+	SegmentsWithCategories int      `json:"segments_with_categories"`
+}
+
+// Run executes artifact-level normalization.
+func Run(ctx context.Context, cfg config.NormalizeConfig) error {
+	if err := ctx.Err(); err != nil {
+		return err
+	}
+
+	parsed, err := ParseFile(cfg.InputFile)
+	if err != nil {
+		return err
+	}
+
+	built, err := Build(parsed, cfg)
+	if err != nil {
+		return err
+	}
+
+	if err := writeOutputJSON(cfg.OutputFile, built.Output); err != nil {
+		return err
+	}
+
+	if cfg.ReportFile != "" {
+		audit := normalizeAudit{
+			Command:                "normalize",
+			InputFile:              cfg.InputFile,
+			OutputFile:             cfg.OutputFile,
+			InputShape:             string(parsed.Shape),
+			InputSegmentCount:      len(parsed.Segments),
+			OutputSchema:           cfg.OutputSchema,
+			OutputModules:          append([]string(nil), cfg.OutputModules...),
+			IDsReassigned:          built.IDsReassigned,
+			SortingChangedInput:    built.SortingChanged,
+			SegmentsWithCategories: built.SegmentsWithCategories,
+		}
+		auditJSON, err := json.Marshal(audit)
+		if err != nil {
+			return fmt.Errorf("marshal normalize audit: %w", err)
+		}
+
+		events := []report.Event{
+			report.Info("normalize", "normalize", "started normalize command"),
+			report.Info("normalize", "normalize", fmt.Sprintf("input file: %s", cfg.InputFile)),
+			report.Info("normalize", "normalize", fmt.Sprintf("detected input shape: %s", parsed.Shape)),
+			report.Info("normalize", "normalize", fmt.Sprintf("input segment count: %d", len(parsed.Segments))),
+			report.Info("normalize", "normalize", fmt.Sprintf("selected output schema: %s", cfg.OutputSchema)),
+			report.Info("normalize", "normalize", fmt.Sprintf("selected output modules: %s", strings.Join(cfg.OutputModules, ","))),
+			report.Info("normalize", "normalize", fmt.Sprintf("output file: %s", cfg.OutputFile)),
+			report.Info("normalize", "normalize", fmt.Sprintf("ids reassigned: %t", built.IDsReassigned)),
+			report.Info("normalize", "normalize", fmt.Sprintf("sorting changed input order: %t", built.SortingChanged)),
+			report.Info("normalize", "normalize", fmt.Sprintf("segments with categories: %d", built.SegmentsWithCategories)),
+			report.Info("normalize", "normalize-audit", string(auditJSON)),
+		}
+		if len(parsed.Segments) == 0 {
+			events = append(events, report.Warning("normalize", "normalize", "input transcript contains zero segments"))
+		}
+		events = append(events,
+			report.Info("normalize", "validate-output", fmt.Sprintf("validated %d output segment(s)", len(parsed.Segments))),
+			report.Info("output", "json", "wrote transcript JSON"),
+		)
+
+		rpt := report.Report{
+			Metadata: report.Metadata{
+				Application:           artifact.ApplicationName,
+				Version:               buildinfo.Version,
+				InputReader:           "normalize-input",
+				InputFiles:            []string{cfg.InputFile},
+				PreprocessingModules:  []string{},
+				PostprocessingModules: []string{},
+				OutputModules:         append([]string(nil), cfg.OutputModules...),
+			},
+			Events: events,
+		}
+		if err := report.WriteJSON(cfg.ReportFile, rpt); err != nil {
+			return fmt.Errorf("write --report-file %q: %w", cfg.ReportFile, err)
+		}
+	}
+
+	return nil
+}
+
+func writeOutputJSON(path string, value any) error {
+	file, err := os.Create(path)
+	if err != nil {
+		return err
+	}
+	defer file.Close()
+
+	encoder := json.NewEncoder(file)
+	encoder.SetIndent("", "  ")
+	if err := encoder.Encode(value); err != nil {
+		return fmt.Errorf("encode normalize output JSON: %w", err)
+	}
+	return nil
+}

internal/normalize/parse.go

@@ -0,0 +1,197 @@
+package normalize
+
+import (
+	"bytes"
+	"encoding/json"
+	"fmt"
+	"io"
+	"os"
+	"strings"
+)
+
+// InputShape identifies which top-level input shape was parsed.
+type InputShape string
+
+const (
+	ShapeObjectWithSegments InputShape = "object_with_segments"
+	ShapeBareSegmentsArray  InputShape = "bare_segments_array"
+)
+
+// ParsedTranscript is the validated normalize input model.
+type ParsedTranscript struct {
+	Shape    InputShape
+	Segments []InputSegment
+}
+
+// InputSegment is a validated segment from normalize input.
+type InputSegment struct {
+	InputIndex         int
+	OriginalID         *int
+	Start              float64
+	End                float64
+	Speaker            string
+	Text               string
+	Categories         []string
+	Source             string
+	SourceSegmentIndex *int
+	SourceRef          string
+	DerivedFrom        []string
+	OverlapGroupID     *int
+}
+
+type inputSegmentPayload struct {
+	ID                 *int     `json:"id"`
+	Start              *float64 `json:"start"`
+	End                *float64 `json:"end"`
+	Speaker            *string  `json:"speaker"`
+	Text               *string  `json:"text"`
+	Categories         []string `json:"categories"`
+	Source             string   `json:"source"`
+	SourceSegmentIndex *int     `json:"source_segment_index"`
+	SourceRef          string   `json:"source_ref"`
+	DerivedFrom        []string `json:"derived_from"`
+	OverlapGroupID     *int     `json:"overlap_group_id"`
+}
+
+// ParseFile parses normalize input JSON from file path.
+func ParseFile(path string) (ParsedTranscript, error) {
+	file, err := os.Open(path)
+	if err != nil {
+		return ParsedTranscript{}, err
+	}
+	defer file.Close()
+
+	return ParseReader(file)
+}
+
+// ParseReader parses normalize input JSON from a reader.
+func ParseReader(reader io.Reader) (ParsedTranscript, error) {
+	var raw json.RawMessage
+	decoder := json.NewDecoder(reader)
+	decoder.UseNumber()
+	if err := decoder.Decode(&raw); err != nil {
+		return ParsedTranscript{}, fmt.Errorf("decode normalize input JSON: %w", err)
+	}
+	if err := ensureSingleValue(decoder); err != nil {
+		return ParsedTranscript{}, err
+	}
+
+	trimmed := bytes.TrimSpace(raw)
+	if len(trimmed) == 0 {
+		return ParsedTranscript{}, fmt.Errorf("normalize input is empty")
+	}
+
+	switch trimmed[0] {
+	case '{':
+		return parseObjectShape(trimmed)
+	case '[':
+		segments, err := parseSegmentsArray(trimmed)
+		if err != nil {
+			return ParsedTranscript{}, err
+		}
+		return ParsedTranscript{
+			Shape:    ShapeBareSegmentsArray,
+			Segments: segments,
+		}, nil
+	default:
+		return ParsedTranscript{}, fmt.Errorf("normalize input must be a top-level object with \"segments\" or a top-level segment array")
+	}
+}
+
+func ensureSingleValue(decoder *json.Decoder) error {
+	var extra json.RawMessage
+	err := decoder.Decode(&extra)
+	if err == io.EOF {
+		return nil
+	}
+	if err == nil {
+		return fmt.Errorf("normalize input must contain exactly one top-level JSON value")
+	}
+	return fmt.Errorf("decode normalize input JSON: %w", err)
+}
+
+func parseObjectShape(raw []byte) (ParsedTranscript, error) {
+	var object map[string]json.RawMessage
+	if err := json.Unmarshal(raw, &object); err != nil {
+		return ParsedTranscript{}, fmt.Errorf("decode normalize object input: %w", err)
+	}
+
+	segmentsRaw, exists := object["segments"]
+	if !exists {
+		return ParsedTranscript{}, fmt.Errorf("normalize object input must contain a \"segments\" field")
+	}
+
+	segments, err := parseSegmentsArray(segmentsRaw)
+	if err != nil {
+		return ParsedTranscript{}, err
+	}
+
+	return ParsedTranscript{
+		Shape:    ShapeObjectWithSegments,
+		Segments: segments,
+	}, nil
+}
+
+func parseSegmentsArray(raw []byte) ([]InputSegment, error) {
+	var segmentValues []json.RawMessage
+	if err := json.Unmarshal(raw, &segmentValues); err != nil {
+		return nil, fmt.Errorf("normalize input \"segments\" must be an array")
+	}
+
+	segments := make([]InputSegment, len(segmentValues))
+	for index, segmentRaw := range segmentValues {
+		segment, err := parseSegment(index, segmentRaw)
+		if err != nil {
+			return nil, err
+		}
+		segments[index] = segment
+	}
+	return segments, nil
+}
+
+func parseSegment(index int, raw []byte) (InputSegment, error) {
+	var payload inputSegmentPayload
+	if err := json.Unmarshal(raw, &payload); err != nil {
+		return InputSegment{}, fmt.Errorf("segment %d: invalid segment object: %w", index, err)
+	}
+
+	if payload.Start == nil {
+		return InputSegment{}, fmt.Errorf("segment %d is missing required field \"start\"", index)
+	}
+	if payload.End == nil {
+		return InputSegment{}, fmt.Errorf("segment %d is missing required field \"end\"", index)
+	}
+	if payload.Speaker == nil {
+		return InputSegment{}, fmt.Errorf("segment %d is missing required field \"speaker\"", index)
+	}
+	if payload.Text == nil {
+		return InputSegment{}, fmt.Errorf("segment %d is missing required field \"text\"", index)
+	}
+
+	if *payload.Start < 0 {
+		return InputSegment{}, fmt.Errorf("segment %d has start %v; start must be >= 0", index, *payload.Start)
+	}
+	if *payload.End < *payload.Start {
+		return InputSegment{}, fmt.Errorf("segment %d has end %v before start %v", index, *payload.End, *payload.Start)
+	}
+
+	speaker := strings.TrimSpace(*payload.Speaker)
+	if speaker == "" {
+		return InputSegment{}, fmt.Errorf("segment %d has empty \"speaker\"; speaker must be non-empty", index)
+	}
+
+	return InputSegment{
+		InputIndex:         index,
+		OriginalID:         payload.ID,
+		Start:              *payload.Start,
+		End:                *payload.End,
+		Speaker:            speaker,
+		Text:               *payload.Text,
+		Categories:         append([]string(nil), payload.Categories...),
+		Source:             payload.Source,
+		SourceSegmentIndex: payload.SourceSegmentIndex,
+		SourceRef:          payload.SourceRef,
+		DerivedFrom:        append([]string(nil), payload.DerivedFrom...),
+		OverlapGroupID:     payload.OverlapGroupID,
+	}, nil
+}

internal/normalize/parse_test.go

@@ -0,0 +1,181 @@
+package normalize
+
+import (
+	"strings"
+	"testing"
+)
+
+func TestParseReaderObjectWithSegmentsParses(t *testing.T) {
+	input := `{
+		"segments": [
+			{"start": 1.0, "end": 2.0, "speaker": " Alice ", "text": "hello", "id": 100}
+		]
+	}`
+
+	parsed, err := ParseReader(strings.NewReader(input))
+	if err != nil {
+		t.Fatalf("parse failed: %v", err)
+	}
+	if parsed.Shape != ShapeObjectWithSegments {
+		t.Fatalf("shape = %q, want %q", parsed.Shape, ShapeObjectWithSegments)
+	}
+	if len(parsed.Segments) != 1 {
+		t.Fatalf("segment count = %d, want 1", len(parsed.Segments))
+	}
+	segment := parsed.Segments[0]
+	if segment.Speaker != "Alice" {
+		t.Fatalf("speaker = %q, want %q", segment.Speaker, "Alice")
+	}
+	if segment.OriginalID == nil || *segment.OriginalID != 100 {
+		t.Fatalf("original id = %v, want 100", segment.OriginalID)
+	}
+}
+
+func TestParseReaderBareSegmentArrayParses(t *testing.T) {
+	input := `[
+		{"start": 1.0, "end": 2.0, "speaker": "Alice", "text": "hello"},
+		{"start": 3.0, "end": 4.0, "speaker": "Bob", "text": "world"}
+	]`
+
+	parsed, err := ParseReader(strings.NewReader(input))
+	if err != nil {
+		t.Fatalf("parse failed: %v", err)
+	}
+	if parsed.Shape != ShapeBareSegmentsArray {
+		t.Fatalf("shape = %q, want %q", parsed.Shape, ShapeBareSegmentsArray)
+	}
+	if len(parsed.Segments) != 2 {
+		t.Fatalf("segment count = %d, want 2", len(parsed.Segments))
+	}
+}
+
+func TestParseReaderInvalidJSONFails(t *testing.T) {
+	_, err := ParseReader(strings.NewReader(`{"segments":`))
+	if err == nil {
+		t.Fatal("expected parse error")
+	}
+	if !strings.Contains(err.Error(), "decode normalize input JSON") {
+		t.Fatalf("unexpected error: %v", err)
+	}
+}
+
+func TestParseReaderObjectMissingSegmentsFails(t *testing.T) {
+	_, err := ParseReader(strings.NewReader(`{"items":[]}`))
+	if err == nil {
+		t.Fatal("expected missing segments error")
+	}
+	if !strings.Contains(err.Error(), "must contain a \"segments\" field") {
+		t.Fatalf("unexpected error: %v", err)
+	}
+}
+
+func TestParseReaderSegmentsNotArrayFails(t *testing.T) {
+	_, err := ParseReader(strings.NewReader(`{"segments": {}}`))
+	if err == nil {
+		t.Fatal("expected segments not array error")
+	}
+	if !strings.Contains(err.Error(), "\"segments\" must be an array") {
+		t.Fatalf("unexpected error: %v", err)
+	}
+}
+
+func TestParseReaderTopLevelScalarShapesFail(t *testing.T) {
+	tests := []string{`"text"`, `42`, `null`, `true`}
+	for _, input := range tests {
+		_, err := ParseReader(strings.NewReader(input))
+		if err == nil {
+			t.Fatalf("expected top-level shape error for %s", input)
+		}
+		if !strings.Contains(err.Error(), "top-level object") {
+			t.Fatalf("unexpected error for %s: %v", input, err)
+		}
+	}
+}
+
+func TestParseReaderMissingStartFails(t *testing.T) {
+	_, err := ParseReader(strings.NewReader(`[{"end":2,"speaker":"A","text":"t"}]`))
+	assertContains(t, err, `missing required field "start"`)
+}
+
+func TestParseReaderMissingEndFails(t *testing.T) {
+	_, err := ParseReader(strings.NewReader(`[{"start":1,"speaker":"A","text":"t"}]`))
+	assertContains(t, err, `missing required field "end"`)
+}
+
+func TestParseReaderMissingSpeakerFails(t *testing.T) {
+	_, err := ParseReader(strings.NewReader(`[{"start":1,"end":2,"text":"t"}]`))
+	assertContains(t, err, `missing required field "speaker"`)
+}
+
+func TestParseReaderEmptySpeakerFails(t *testing.T) {
+	_, err := ParseReader(strings.NewReader(`[{"start":1,"end":2,"speaker":"   ","text":"t"}]`))
+	assertContains(t, err, `speaker must be non-empty`)
+}
+
+func TestParseReaderMissingTextFails(t *testing.T) {
+	_, err := ParseReader(strings.NewReader(`[{"start":1,"end":2,"speaker":"A"}]`))
+	assertContains(t, err, `missing required field "text"`)
+}
+
+func TestParseReaderEndBeforeStartFails(t *testing.T) {
+	_, err := ParseReader(strings.NewReader(`[{"start":3,"end":2,"speaker":"A","text":"t"}]`))
+	assertContains(t, err, "before start")
+}
+
+func TestParseReaderNegativeStartFails(t *testing.T) {
+	_, err := ParseReader(strings.NewReader(`[{"start":-1,"end":2,"speaker":"A","text":"t"}]`))
+	assertContains(t, err, "start must be >= 0")
+}
+
+func TestParseReaderEmptySegmentsArrayAccepted(t *testing.T) {
+	parsed, err := ParseReader(strings.NewReader(`{"segments":[]}`))
+	if err != nil {
+		t.Fatalf("parse failed: %v", err)
+	}
+	if len(parsed.Segments) != 0 {
+		t.Fatalf("segment count = %d, want 0", len(parsed.Segments))
+	}
+}
+
+func TestParseReaderCategoriesPreservedWhenValid(t *testing.T) {
+	parsed, err := ParseReader(strings.NewReader(`[{"start":1,"end":2,"speaker":"A","text":"t","categories":["filler","backchannel"]}]`))
+	if err != nil {
+		t.Fatalf("parse failed: %v", err)
+	}
+	if len(parsed.Segments) != 1 {
+		t.Fatalf("segment count = %d, want 1", len(parsed.Segments))
+	}
+	if len(parsed.Segments[0].Categories) != 2 {
+		t.Fatalf("categories length = %d, want 2", len(parsed.Segments[0].Categories))
+	}
+	if parsed.Segments[0].Categories[0] != "filler" || parsed.Segments[0].Categories[1] != "backchannel" {
+		t.Fatalf("categories = %v", parsed.Segments[0].Categories)
+	}
+}
+
+func TestParseReaderOriginalInputIndexPreserved(t *testing.T) {
+	input := `[
+		{"start":1,"end":2,"speaker":"A","text":"one"},
+		{"start":2,"end":3,"speaker":"B","text":"two"},
+		{"start":3,"end":4,"speaker":"C","text":"three"}
+	]`
+	parsed, err := ParseReader(strings.NewReader(input))
+	if err != nil {
+		t.Fatalf("parse failed: %v", err)
+	}
+	for index, segment := range parsed.Segments {
+		if segment.InputIndex != index {
+			t.Fatalf("segment %d input index = %d, want %d", index, segment.InputIndex, index)
+		}
+	}
+}
+
+func assertContains(t *testing.T, err error, fragment string) {
+	t.Helper()
+	if err == nil {
+		t.Fatalf("expected error containing %q", fragment)
+	}
+	if !strings.Contains(err.Error(), fragment) {
+		t.Fatalf("error = %q, want substring %q", err.Error(), fragment)
+	}
+}

internal/trim/apply.go

@@ -0,0 +1,367 @@
+package trim
+
+import (
+	"fmt"
+
+	"gitea.maximumdirect.net/eric/seriatim/internal/model"
+	"gitea.maximumdirect.net/eric/seriatim/internal/overlap"
+	"gitea.maximumdirect.net/eric/seriatim/schema"
+)
+
+// Mode controls how selector IDs are applied.
+type Mode string
+
+const (
+	ModeKeep   Mode = "keep"
+	ModeRemove Mode = "remove"
+)
+
+// Options configures transcript trimming.
+type Options struct {
+	Mode       Mode
+	Selector   Selector
+	AllowEmpty bool
+}
+
+// Result contains trimming output and ID mapping metadata.
+type Result struct {
+	Transcript schema.Transcript
+	OldToNewID map[int]int
+	RemovedIDs []int
+}
+
+// IntermediateResult contains trimming output for intermediate schema artifacts.
+type IntermediateResult struct {
+	Transcript schema.IntermediateTranscript
+	OldToNewID map[int]int
+	RemovedIDs []int
+}
+
+// MinimalResult contains trimming output for minimal schema artifacts.
+type MinimalResult struct {
+	Transcript schema.MinimalTranscript
+	OldToNewID map[int]int
+	RemovedIDs []int
+}
+
+// Apply trims a full seriatim output transcript by segment ID.
+func Apply(input schema.Transcript, opts Options) (Result, error) {
+	if err := validateMode(opts.Mode); err != nil {
+		return Result{}, err
+	}
+
+	selected := opts.Selector.IDs()
+	if len(selected) == 0 {
+		return Result{}, fmt.Errorf("selector cannot be empty")
+	}
+
+	inputIDs := make([]int, len(input.Segments))
+	for index, segment := range input.Segments {
+		inputIDs[index] = segment.ID
+	}
+
+	idIndex, err := validateInputIDs(inputIDs)
+	if err != nil {
+		return Result{}, err
+	}
+
+	if err := validateSelectedIDsExist(selected, idIndex); err != nil {
+		return Result{}, err
+	}
+
+	kept := make([]schema.Segment, 0, len(input.Segments))
+	removed := make([]int, 0, len(input.Segments))
+	oldToNew := make(map[int]int, len(input.Segments))
+	for _, segment := range input.Segments {
+		keep := opts.Mode == ModeKeep && opts.Selector.Contains(segment.ID)
+		if opts.Mode == ModeRemove {
+			keep = !opts.Selector.Contains(segment.ID)
+		}
+
+		if !keep {
+			removed = append(removed, segment.ID)
+			continue
+		}
+
+		rewritten := copySegment(segment)
+		rewritten.ID = len(kept) + 1
+		rewritten.OverlapGroupID = 0
+		kept = append(kept, rewritten)
+		oldToNew[segment.ID] = rewritten.ID
+	}
+
+	if len(kept) == 0 && !opts.AllowEmpty {
+		return Result{}, fmt.Errorf("trim operation produced an empty transcript; set AllowEmpty to true to permit this")
+	}
+
+	kept, groups := recomputeOverlapGroups(kept)
+	if groups == nil {
+		groups = make([]schema.OverlapGroup, 0)
+	}
+
+	out := copyTranscript(input)
+	out.Segments = kept
+	out.OverlapGroups = groups
+	return Result{
+		Transcript: out,
+		OldToNewID: oldToNew,
+		RemovedIDs: removed,
+	}, nil
+}
+
+// ApplyIntermediate trims an intermediate seriatim output transcript by
+// segment ID.
+func ApplyIntermediate(input schema.IntermediateTranscript, opts Options) (IntermediateResult, error) {
+	if err := validateMode(opts.Mode); err != nil {
+		return IntermediateResult{}, err
+	}
+
+	selected := opts.Selector.IDs()
+	if len(selected) == 0 {
+		return IntermediateResult{}, fmt.Errorf("selector cannot be empty")
+	}
+
+	inputIDs := make([]int, len(input.Segments))
+	for index, segment := range input.Segments {
+		inputIDs[index] = segment.ID
+	}
+	idIndex, err := validateInputIDs(inputIDs)
+	if err != nil {
+		return IntermediateResult{}, err
+	}
+	if err := validateSelectedIDsExist(selected, idIndex); err != nil {
+		return IntermediateResult{}, err
+	}
+
+	kept := make([]schema.IntermediateSegment, 0, len(input.Segments))
+	removed := make([]int, 0, len(input.Segments))
+	oldToNew := make(map[int]int, len(input.Segments))
+	for _, segment := range input.Segments {
+		keep := opts.Mode == ModeKeep && opts.Selector.Contains(segment.ID)
+		if opts.Mode == ModeRemove {
+			keep = !opts.Selector.Contains(segment.ID)
+		}
+		if !keep {
+			removed = append(removed, segment.ID)
+			continue
+		}
+
+		rewritten := schema.IntermediateSegment{
+			ID:         len(kept) + 1,
+			Start:      segment.Start,
+			End:        segment.End,
+			Speaker:    segment.Speaker,
+			Text:       segment.Text,
+			Categories: append([]string(nil), segment.Categories...),
+		}
+		kept = append(kept, rewritten)
+		oldToNew[segment.ID] = rewritten.ID
+	}
+
+	if len(kept) == 0 && !opts.AllowEmpty {
+		return IntermediateResult{}, fmt.Errorf("trim operation produced an empty transcript; set AllowEmpty to true to permit this")
+	}
+
+	return IntermediateResult{
+		Transcript: schema.IntermediateTranscript{
+			Metadata: schema.IntermediateMetadata{
+				Application:  input.Metadata.Application,
+				Version:      input.Metadata.Version,
+				OutputSchema: input.Metadata.OutputSchema,
+			},
+			Segments: kept,
+		},
+		OldToNewID: oldToNew,
+		RemovedIDs: removed,
+	}, nil
+}
+
+// ApplyMinimal trims a minimal seriatim output transcript by segment ID.
+func ApplyMinimal(input schema.MinimalTranscript, opts Options) (MinimalResult, error) {
+	if err := validateMode(opts.Mode); err != nil {
+		return MinimalResult{}, err
+	}
+
+	selected := opts.Selector.IDs()
+	if len(selected) == 0 {
+		return MinimalResult{}, fmt.Errorf("selector cannot be empty")
+	}
+
+	inputIDs := make([]int, len(input.Segments))
+	for index, segment := range input.Segments {
+		inputIDs[index] = segment.ID
+	}
+	idIndex, err := validateInputIDs(inputIDs)
+	if err != nil {
+		return MinimalResult{}, err
+	}
+	if err := validateSelectedIDsExist(selected, idIndex); err != nil {
+		return MinimalResult{}, err
+	}
+
+	kept := make([]schema.MinimalSegment, 0, len(input.Segments))
+	removed := make([]int, 0, len(input.Segments))
+	oldToNew := make(map[int]int, len(input.Segments))
+	for _, segment := range input.Segments {
+		keep := opts.Mode == ModeKeep && opts.Selector.Contains(segment.ID)
+		if opts.Mode == ModeRemove {
+			keep = !opts.Selector.Contains(segment.ID)
+		}
+		if !keep {
+			removed = append(removed, segment.ID)
+			continue
+		}
+
+		rewritten := schema.MinimalSegment{
+			ID:      len(kept) + 1,
+			Start:   segment.Start,
+			End:     segment.End,
+			Speaker: segment.Speaker,
+			Text:    segment.Text,
+		}
+		kept = append(kept, rewritten)
+		oldToNew[segment.ID] = rewritten.ID
+	}
+
+	if len(kept) == 0 && !opts.AllowEmpty {
+		return MinimalResult{}, fmt.Errorf("trim operation produced an empty transcript; set AllowEmpty to true to permit this")
+	}
+
+	return MinimalResult{
+		Transcript: schema.MinimalTranscript{
+			Metadata: schema.MinimalMetadata{
+				Application:  input.Metadata.Application,
+				Version:      input.Metadata.Version,
+				OutputSchema: input.Metadata.OutputSchema,
+			},
+			Segments: kept,
+		},
+		OldToNewID: oldToNew,
+		RemovedIDs: removed,
+	}, nil
+}
+
+func validateMode(mode Mode) error {
+	switch mode {
+	case ModeKeep, ModeRemove:
+		return nil
+	default:
+		return fmt.Errorf("invalid trim mode %q", mode)
+	}
+}
+
+func validateInputIDs(ids []int) (map[int]int, error) {
+	seen := make(map[int]int, len(ids))
+	for index, id := range ids {
+		if id <= 0 {
+			return nil, fmt.Errorf("input transcript has non-positive segment ID %d at index %d", id, index)
+		}
+		if firstIndex, exists := seen[id]; exists {
+			return nil, fmt.Errorf("input transcript has duplicate segment ID %d at indexes %d and %d", id, firstIndex, index)
+		}
+		seen[id] = index
+	}
+
+	for id := 1; id <= len(ids); id++ {
+		if _, exists := seen[id]; !exists {
+			return nil, fmt.Errorf("input transcript segment IDs must be sequential 1..%d; missing ID %d", len(ids), id)
+		}
+	}
+	return seen, nil
+}
+
+func validateSelectedIDsExist(selected []int, idIndex map[int]int) error {
+	for _, id := range selected {
+		if _, exists := idIndex[id]; !exists {
+			return fmt.Errorf("selected segment ID %d does not exist in input transcript", id)
+		}
+	}
+	return nil
+}
+
+func recomputeOverlapGroups(segments []schema.Segment) ([]schema.Segment, []schema.OverlapGroup) {
+	if len(segments) == 0 {
+		return segments, make([]schema.OverlapGroup, 0)
+	}
+
+	modelSegments := make([]model.Segment, len(segments))
+	for index, segment := range segments {
+		modelSegments[index] = model.Segment{
+			ID:                 segment.ID,
+			Source:             segment.Source,
+			SourceSegmentIndex: copyIntPtr(segment.SourceSegmentIndex),
+			SourceRef:          segment.SourceRef,
+			DerivedFrom:        append([]string(nil), segment.DerivedFrom...),
+			Speaker:            segment.Speaker,
+			Start:              segment.Start,
+			End:                segment.End,
+			Text:               segment.Text,
+			Categories:         append([]string(nil), segment.Categories...),
+			OverlapGroupID:     segment.OverlapGroupID,
+		}
+	}
+
+	detected := overlap.Detect(model.MergedTranscript{
+		Segments: modelSegments,
+	})
+	rewrittenSegments := make([]schema.Segment, len(segments))
+	for index, segment := range segments {
+		rewritten := copySegment(segment)
+		rewritten.OverlapGroupID = detected.Segments[index].OverlapGroupID
+		rewrittenSegments[index] = rewritten
+	}
+
+	groups := make([]schema.OverlapGroup, len(detected.OverlapGroups))
+	for index, group := range detected.OverlapGroups {
+		groups[index] = schema.OverlapGroup{
+			ID:         group.ID,
+			Start:      group.Start,
+			End:        group.End,
+			Segments:   append([]string(nil), group.Segments...),
+			Speakers:   append([]string(nil), group.Speakers...),
+			Class:      group.Class,
+			Resolution: group.Resolution,
+		}
+	}
+	return rewrittenSegments, groups
+}
+
+func copyTranscript(input schema.Transcript) schema.Transcript {
+	return schema.Transcript{
+		Metadata: schema.Metadata{
+			Application:           input.Metadata.Application,
+			Version:               input.Metadata.Version,
+			InputReader:           input.Metadata.InputReader,
+			InputFiles:            append([]string(nil), input.Metadata.InputFiles...),
+			PreprocessingModules:  append([]string(nil), input.Metadata.PreprocessingModules...),
+			PostprocessingModules: append([]string(nil), input.Metadata.PostprocessingModules...),
+			OutputModules:         append([]string(nil), input.Metadata.OutputModules...),
+		},
+		Segments:      append([]schema.Segment(nil), input.Segments...),
+		OverlapGroups: append([]schema.OverlapGroup(nil), input.OverlapGroups...),
+	}
+}
+
+func copySegment(input schema.Segment) schema.Segment {
+	return schema.Segment{
+		ID:                 input.ID,
+		Source:             input.Source,
+		SourceSegmentIndex: copyIntPtr(input.SourceSegmentIndex),
+		SourceRef:          input.SourceRef,
+		DerivedFrom:        append([]string(nil), input.DerivedFrom...),
+		Speaker:            input.Speaker,
+		Start:              input.Start,
+		End:                input.End,
+		Text:               input.Text,
+		Categories:         append([]string(nil), input.Categories...),
+		OverlapGroupID:     input.OverlapGroupID,
+	}
+}
+
+func copyIntPtr(value *int) *int {
+	if value == nil {
+		return nil
+	}
+	copied := *value
+	return &copied
+}

internal/trim/apply_test.go

@@ -0,0 +1,668 @@
+package trim
+
+import (
+	"strings"
+	"testing"
+
+	"gitea.maximumdirect.net/eric/seriatim/schema"
+)
+
+func TestApplyKeepModeRenumbersFromOne(t *testing.T) {
+	input := fullTranscriptFixture()
+	selector := mustParseSelector(t, "2,4")
+
+	result, err := Apply(input, Options{
+		Mode:     ModeKeep,
+		Selector: selector,
+	})
+	if err != nil {
+		t.Fatalf("apply failed: %v", err)
+	}
+
+	if len(result.Transcript.Segments) != 2 {
+		t.Fatalf("segment count = %d, want 2", len(result.Transcript.Segments))
+	}
+	assertSegmentIDs(t, result.Transcript.Segments, []int{1, 2})
+	assertSegmentTexts(t, result.Transcript.Segments, []string{"beta", "delta"})
+	assertIntMap(t, result.OldToNewID, map[int]int{2: 1, 4: 2})
+	assertIntSlice(t, result.RemovedIDs, []int{1, 3})
+}
+
+func TestApplyRemoveModeRenumbersFromOne(t *testing.T) {
+	input := fullTranscriptFixture()
+	selector := mustParseSelector(t, "2,4")
+
+	result, err := Apply(input, Options{
+		Mode:     ModeRemove,
+		Selector: selector,
+	})
+	if err != nil {
+		t.Fatalf("apply failed: %v", err)
+	}
+
+	assertSegmentIDs(t, result.Transcript.Segments, []int{1, 2})
+	assertSegmentTexts(t, result.Transcript.Segments, []string{"alpha", "gamma"})
+	assertIntMap(t, result.OldToNewID, map[int]int{1: 1, 3: 2})
+	assertIntSlice(t, result.RemovedIDs, []int{2, 4})
+}
+
+func TestApplySelectorOrderDoesNotChangeTranscriptOrder(t *testing.T) {
+	input := fullTranscriptFixture()
+	selector := mustParseSelector(t, "4,1,3")
+
+	result, err := Apply(input, Options{
+		Mode:     ModeKeep,
+		Selector: selector,
+	})
+	if err != nil {
+		t.Fatalf("apply failed: %v", err)
+	}
+
+	assertSegmentIDs(t, result.Transcript.Segments, []int{1, 2, 3})
+	assertSegmentTexts(t, result.Transcript.Segments, []string{"alpha", "gamma", "delta"})
+}
+
+func TestApplyFailsWhenSelectedIDDoesNotExist(t *testing.T) {
+	input := fullTranscriptFixture()
+	selector := mustParseSelector(t, "2,99")
+
+	_, err := Apply(input, Options{
+		Mode:     ModeKeep,
+		Selector: selector,
+	})
+	if err == nil {
+		t.Fatal("expected missing selected ID error")
+	}
+	if !strings.Contains(err.Error(), "does not exist") {
+		t.Fatalf("unexpected error: %v", err)
+	}
+}
+
+func TestApplyFailsOnDuplicateInputIDs(t *testing.T) {
+	input := fullTranscriptFixture()
+	input.Segments[2].ID = 2
+	selector := mustParseSelector(t, "2")
+
+	_, err := Apply(input, Options{
+		Mode:     ModeKeep,
+		Selector: selector,
+	})
+	if err == nil {
+		t.Fatal("expected duplicate input ID error")
+	}
+	if !strings.Contains(err.Error(), "duplicate segment ID") {
+		t.Fatalf("unexpected error: %v", err)
+	}
+}
+
+func TestApplyFailsOnMissingOrNonSequentialInputIDs(t *testing.T) {
+	input := fullTranscriptFixture()
+	input.Segments[1].ID = 5
+	selector := mustParseSelector(t, "1")
+
+	_, err := Apply(input, Options{
+		Mode:     ModeKeep,
+		Selector: selector,
+	})
+	if err == nil {
+		t.Fatal("expected non-sequential input ID error")
+	}
+	if !strings.Contains(err.Error(), "must be sequential") {
+		t.Fatalf("unexpected error: %v", err)
+	}
+}
+
+func TestApplyFailsOnNonPositiveInputIDs(t *testing.T) {
+	input := fullTranscriptFixture()
+	input.Segments[0].ID = 0
+	selector := mustParseSelector(t, "1")
+
+	_, err := Apply(input, Options{
+		Mode:     ModeKeep,
+		Selector: selector,
+	})
+	if err == nil {
+		t.Fatal("expected non-positive input ID error")
+	}
+	if !strings.Contains(err.Error(), "non-positive") {
+		t.Fatalf("unexpected error: %v", err)
+	}
+}
+
+func TestApplyEmptyOutputFailsUnlessAllowEmpty(t *testing.T) {
+	input := fullTranscriptFixture()
+	selector := mustParseSelector(t, "1-4")
+
+	_, err := Apply(input, Options{
+		Mode:     ModeRemove,
+		Selector: selector,
+	})
+	if err == nil {
+		t.Fatal("expected empty-output error")
+	}
+	if !strings.Contains(err.Error(), "empty transcript") {
+		t.Fatalf("unexpected error: %v", err)
+	}
+
+	allowed, err := Apply(input, Options{
+		Mode:       ModeRemove,
+		Selector:   selector,
+		AllowEmpty: true,
+	})
+	if err != nil {
+		t.Fatalf("apply with AllowEmpty failed: %v", err)
+	}
+	if len(allowed.Transcript.Segments) != 0 {
+		t.Fatalf("segment count = %d, want 0", len(allowed.Transcript.Segments))
+	}
+	assertIntMap(t, allowed.OldToNewID, map[int]int{})
+	assertIntSlice(t, allowed.RemovedIDs, []int{1, 2, 3, 4})
+}
+
+func TestApplyPreservesRetainedSegmentFieldsAndClearsOverlapIDs(t *testing.T) {
+	input := fullTranscriptFixture()
+	selector := mustParseSelector(t, "2")
+
+	result, err := Apply(input, Options{
+		Mode:     ModeKeep,
+		Selector: selector,
+	})
+	if err != nil {
+		t.Fatalf("apply failed: %v", err)
+	}
+
+	if len(result.Transcript.Segments) != 1 {
+		t.Fatalf("segment count = %d, want 1", len(result.Transcript.Segments))
+	}
+	segment := result.Transcript.Segments[0]
+	if segment.ID != 1 {
+		t.Fatalf("segment ID = %d, want 1", segment.ID)
+	}
+	if segment.Source != "b.json" {
+		t.Fatalf("source = %q, want %q", segment.Source, "b.json")
+	}
+	if segment.SourceSegmentIndex == nil || *segment.SourceSegmentIndex != 20 {
+		t.Fatalf("source_segment_index = %v, want 20", segment.SourceSegmentIndex)
+	}
+	if segment.SourceRef != "b.json#20" {
+		t.Fatalf("source_ref = %q, want %q", segment.SourceRef, "b.json#20")
+	}
+	if !equalStringSlices(segment.DerivedFrom, []string{"b.json#19", "b.json#20"}) {
+		t.Fatalf("derived_from = %v, want %v", segment.DerivedFrom, []string{"b.json#19", "b.json#20"})
+	}
+	if !equalStringSlices(segment.Categories, []string{"filler", "backchannel"}) {
+		t.Fatalf("categories = %v, want %v", segment.Categories, []string{"filler", "backchannel"})
+	}
+	if segment.Speaker != "Bob" {
+		t.Fatalf("speaker = %q, want Bob", segment.Speaker)
+	}
+	if segment.Start != 2 || segment.End != 3 {
+		t.Fatalf("times = %.3f-%.3f, want 2.000-3.000", segment.Start, segment.End)
+	}
+	if segment.Text != "beta" {
+		t.Fatalf("text = %q, want beta", segment.Text)
+	}
+	if segment.OverlapGroupID != 0 {
+		t.Fatalf("overlap_group_id = %d, want 0", segment.OverlapGroupID)
+	}
+	if len(result.Transcript.OverlapGroups) != 0 {
+		t.Fatalf("overlap_groups count = %d, want 0", len(result.Transcript.OverlapGroups))
+	}
+}
+
+func TestApplyFullSchemaRemovesStaleOverlapGroups(t *testing.T) {
+	input := overlapTranscriptFixture()
+	selector := mustParseSelector(t, "1,3")
+
+	result, err := Apply(input, Options{
+		Mode:     ModeKeep,
+		Selector: selector,
+	})
+	if err != nil {
+		t.Fatalf("apply failed: %v", err)
+	}
+
+	if len(result.Transcript.OverlapGroups) != 0 {
+		t.Fatalf("overlap_groups count = %d, want 0", len(result.Transcript.OverlapGroups))
+	}
+	for index, segment := range result.Transcript.Segments {
+		if segment.OverlapGroupID != 0 {
+			t.Fatalf("segment %d overlap_group_id = %d, want 0", index, segment.OverlapGroupID)
+		}
+	}
+}
+
+func TestApplyFullSchemaRecomputesOverlapGroup(t *testing.T) {
+	input := overlapTranscriptFixture()
+	selector := mustParseSelector(t, "1,2")
+
+	result, err := Apply(input, Options{
+		Mode:     ModeKeep,
+		Selector: selector,
+	})
+	if err != nil {
+		t.Fatalf("apply failed: %v", err)
+	}
+
+	assertSegmentIDs(t, result.Transcript.Segments, []int{1, 2})
+	assertIntSlice(t, []int{
+		result.Transcript.Segments[0].OverlapGroupID,
+		result.Transcript.Segments[1].OverlapGroupID,
+	}, []int{1, 1})
+	if len(result.Transcript.OverlapGroups) != 1 {
+		t.Fatalf("overlap_groups count = %d, want 1", len(result.Transcript.OverlapGroups))
+	}
+	group := result.Transcript.OverlapGroups[0]
+	if group.ID != 1 {
+		t.Fatalf("group ID = %d, want 1", group.ID)
+	}
+	if group.Start != 1 || group.End != 4 {
+		t.Fatalf("group times = %.3f-%.3f, want 1.000-4.000", group.Start, group.End)
+	}
+	if !equalStringSlices(group.Segments, []string{"a.json#10", "b.json#20"}) {
+		t.Fatalf("group segments = %v, want %v", group.Segments, []string{"a.json#10", "b.json#20"})
+	}
+	if !equalStringSlices(group.Speakers, []string{"Alice", "Bob"}) {
+		t.Fatalf("group speakers = %v, want %v", group.Speakers, []string{"Alice", "Bob"})
+	}
+}
+
+func TestApplyFullSchemaDropsGroupWhenFewerThanTwoSpeakersRemain(t *testing.T) {
+	input := overlapTranscriptFixture()
+	selector := mustParseSelector(t, "1")
+
+	result, err := Apply(input, Options{
+		Mode:     ModeKeep,
+		Selector: selector,
+	})
+	if err != nil {
+		t.Fatalf("apply failed: %v", err)
+	}
+
+	if len(result.Transcript.OverlapGroups) != 0 {
+		t.Fatalf("overlap_groups count = %d, want 0", len(result.Transcript.OverlapGroups))
+	}
+	if len(result.Transcript.Segments) != 1 {
+		t.Fatalf("segment count = %d, want 1", len(result.Transcript.Segments))
+	}
+	if result.Transcript.Segments[0].OverlapGroupID != 0 {
+		t.Fatalf("segment overlap_group_id = %d, want 0", result.Transcript.Segments[0].OverlapGroupID)
+	}
+}
+
+func TestApplyFullSchemaHandlesTransitiveOverlaps(t *testing.T) {
+	input := transitiveOverlapFixture()
+	selector := mustParseSelector(t, "1-3")
+
+	result, err := Apply(input, Options{
+		Mode:     ModeKeep,
+		Selector: selector,
+	})
+	if err != nil {
+		t.Fatalf("apply failed: %v", err)
+	}
+
+	if len(result.Transcript.OverlapGroups) != 1 {
+		t.Fatalf("overlap_groups count = %d, want 1", len(result.Transcript.OverlapGroups))
+	}
+	assertIntSlice(t, []int{
+		result.Transcript.Segments[0].OverlapGroupID,
+		result.Transcript.Segments[1].OverlapGroupID,
+		result.Transcript.Segments[2].OverlapGroupID,
+	}, []int{1, 1, 1})
+	group := result.Transcript.OverlapGroups[0]
+	if group.Start != 10 || group.End != 15 {
+		t.Fatalf("group times = %.3f-%.3f, want 10.000-15.000", group.Start, group.End)
+	}
+}
+
+func TestApplyFullSchemaBoundaryTouchingNotGrouped(t *testing.T) {
+	input := boundaryFixture()
+	selector := mustParseSelector(t, "1-2")
+
+	result, err := Apply(input, Options{
+		Mode:     ModeKeep,
+		Selector: selector,
+	})
+	if err != nil {
+		t.Fatalf("apply failed: %v", err)
+	}
+
+	if len(result.Transcript.OverlapGroups) != 0 {
+		t.Fatalf("overlap_groups count = %d, want 0", len(result.Transcript.OverlapGroups))
+	}
+	assertIntSlice(t, []int{
+		result.Transcript.Segments[0].OverlapGroupID,
+		result.Transcript.Segments[1].OverlapGroupID,
+	}, []int{0, 0})
+}
+
+func TestApplyIntermediateDoesNotIncludeOverlapGroups(t *testing.T) {
+	input := schema.IntermediateTranscript{
+		Metadata: schema.IntermediateMetadata{
+			Application:  "seriatim",
+			Version:      "v-test",
+			OutputSchema: "seriatim-intermediate",
+		},
+		Segments: []schema.IntermediateSegment{
+			{ID: 1, Start: 1, End: 3, Speaker: "Alice", Text: "alpha", Categories: []string{"word-run"}},
+			{ID: 2, Start: 2, End: 4, Speaker: "Bob", Text: "beta", Categories: []string{"filler"}},
+		},
+	}
+	selector := mustParseSelector(t, "1")
+	result, err := ApplyIntermediate(input, Options{
+		Mode:     ModeKeep,
+		Selector: selector,
+	})
+	if err != nil {
+		t.Fatalf("apply intermediate failed: %v", err)
+	}
+	if len(result.Transcript.Segments) != 1 {
+		t.Fatalf("segment count = %d, want 1", len(result.Transcript.Segments))
+	}
+	if result.Transcript.Segments[0].ID != 1 {
+		t.Fatalf("segment id = %d, want 1", result.Transcript.Segments[0].ID)
+	}
+	if err := schema.ValidateIntermediateTranscript(result.Transcript); err != nil {
+		t.Fatalf("intermediate output should remain valid: %v", err)
+	}
+}
+
+func TestApplyMinimalDoesNotIncludeOverlapGroups(t *testing.T) {
+	input := schema.MinimalTranscript{
+		Metadata: schema.MinimalMetadata{
+			Application:  "seriatim",
+			Version:      "v-test",
+			OutputSchema: "seriatim-minimal",
+		},
+		Segments: []schema.MinimalSegment{
+			{ID: 1, Start: 1, End: 3, Speaker: "Alice", Text: "alpha"},
+			{ID: 2, Start: 2, End: 4, Speaker: "Bob", Text: "beta"},
+		},
+	}
+	selector := mustParseSelector(t, "2")
+	result, err := ApplyMinimal(input, Options{
+		Mode:     ModeKeep,
+		Selector: selector,
+	})
+	if err != nil {
+		t.Fatalf("apply minimal failed: %v", err)
+	}
+	if len(result.Transcript.Segments) != 1 {
+		t.Fatalf("segment count = %d, want 1", len(result.Transcript.Segments))
+	}
+	if result.Transcript.Segments[0].ID != 1 {
+		t.Fatalf("segment id = %d, want 1", result.Transcript.Segments[0].ID)
+	}
+	if err := schema.ValidateMinimalTranscript(result.Transcript); err != nil {
+		t.Fatalf("minimal output should remain valid: %v", err)
+	}
+}
+
+func TestApplyOutputInvariantsValidAfterRenumberAndOverlapRecompute(t *testing.T) {
+	input := overlapTranscriptFixture()
+	selector := mustParseSelector(t, "2,1")
+
+	result, err := Apply(input, Options{
+		Mode:     ModeKeep,
+		Selector: selector,
+	})
+	if err != nil {
+		t.Fatalf("apply failed: %v", err)
+	}
+
+	if err := schema.ValidateTranscript(result.Transcript); err != nil {
+		t.Fatalf("trim output should remain valid: %v", err)
+	}
+}
+
+func mustParseSelector(t *testing.T, value string) Selector {
+	t.Helper()
+	selector, err := ParseSelector(value)
+	if err != nil {
+		t.Fatalf("selector parse failed for %q: %v", value, err)
+	}
+	return selector
+}
+
+func fullTranscriptFixture() schema.Transcript {
+	firstIndex := 10
+	secondIndex := 20
+	thirdIndex := 30
+	fourthIndex := 40
+
+	return schema.Transcript{
+		Metadata: schema.Metadata{
+			Application:           "seriatim",
+			Version:               "v-test",
+			InputReader:           "json-files",
+			InputFiles:            []string{"a.json", "b.json"},
+			PreprocessingModules:  []string{"validate-raw"},
+			PostprocessingModules: []string{"detect-overlaps"},
+			OutputModules:         []string{"json"},
+		},
+		Segments: []schema.Segment{
+			{
+				ID:                 1,
+				Source:             "a.json",
+				SourceSegmentIndex: &firstIndex,
+				SourceRef:          "a.json#10",
+				DerivedFrom:        []string{"a.json#10"},
+				Speaker:            "Alice",
+				Start:              1,
+				End:                2,
+				Text:               "alpha",
+				Categories:         []string{"word-run"},
+				OverlapGroupID:     7,
+			},
+			{
+				ID:                 2,
+				Source:             "b.json",
+				SourceSegmentIndex: &secondIndex,
+				SourceRef:          "b.json#20",
+				DerivedFrom:        []string{"b.json#19", "b.json#20"},
+				Speaker:            "Bob",
+				Start:              2,
+				End:                3,
+				Text:               "beta",
+				Categories:         []string{"filler", "backchannel"},
+				OverlapGroupID:     7,
+			},
+			{
+				ID:                 3,
+				Source:             "c.json",
+				SourceSegmentIndex: &thirdIndex,
+				SourceRef:          "c.json#30",
+				DerivedFrom:        []string{"c.json#30"},
+				Speaker:            "Carol",
+				Start:              3,
+				End:                4,
+				Text:               "gamma",
+				Categories:         []string{"normal"},
+				OverlapGroupID:     8,
+			},
+			{
+				ID:                 4,
+				Source:             "d.json",
+				SourceSegmentIndex: &fourthIndex,
+				SourceRef:          "d.json#40",
+				DerivedFrom:        []string{"d.json#40"},
+				Speaker:            "Dan",
+				Start:              4,
+				End:                5,
+				Text:               "delta",
+				Categories:         []string{"normal"},
+				OverlapGroupID:     9,
+			},
+		},
+		OverlapGroups: []schema.OverlapGroup{
+			{
+				ID:         7,
+				Start:      1.5,
+				End:        3.1,
+				Segments:   []string{"a.json#10", "b.json#20"},
+				Speakers:   []string{"Alice", "Bob"},
+				Class:      "unknown",
+				Resolution: "unresolved",
+			},
+		},
+	}
+}
+
+func overlapTranscriptFixture() schema.Transcript {
+	first := 10
+	second := 20
+	third := 30
+
+	return schema.Transcript{
+		Metadata: schema.Metadata{
+			Application:           "seriatim",
+			Version:               "v-test",
+			InputReader:           "json-files",
+			InputFiles:            []string{"a.json", "b.json", "c.json"},
+			PreprocessingModules:  []string{"validate-raw"},
+			PostprocessingModules: []string{"detect-overlaps"},
+			OutputModules:         []string{"json"},
+		},
+		Segments: []schema.Segment{
+			{
+				ID:                 1,
+				Source:             "a.json",
+				SourceSegmentIndex: &first,
+				SourceRef:          "a.json#10",
+				Speaker:            "Alice",
+				Start:              1,
+				End:                4,
+				Text:               "a",
+				OverlapGroupID:     99,
+			},
+			{
+				ID:                 2,
+				Source:             "b.json",
+				SourceSegmentIndex: &second,
+				SourceRef:          "b.json#20",
+				Speaker:            "Bob",
+				Start:              2,
+				End:                3,
+				Text:               "b",
+				OverlapGroupID:     99,
+			},
+			{
+				ID:                 3,
+				Source:             "c.json",
+				SourceSegmentIndex: &third,
+				SourceRef:          "c.json#30",
+				Speaker:            "Carol",
+				Start:              10,
+				End:                11,
+				Text:               "c",
+				OverlapGroupID:     100,
+			},
+		},
+		OverlapGroups: []schema.OverlapGroup{
+			{
+				ID:         99,
+				Start:      0,
+				End:        100,
+				Segments:   []string{"stale#1", "stale#2"},
+				Speakers:   []string{"stale"},
+				Class:      "unknown",
+				Resolution: "unresolved",
+			},
+		},
+	}
+}
+
+func transitiveOverlapFixture() schema.Transcript {
+	one := 1
+	two := 2
+	three := 3
+	return schema.Transcript{
+		Metadata: schema.Metadata{
+			Application: "seriatim",
+			Version:     "v-test",
+		},
+		Segments: []schema.Segment{
+			{ID: 1, Source: "a.json", SourceSegmentIndex: &one, Speaker: "Alice", Start: 10, End: 14, Text: "a"},
+			{ID: 2, Source: "b.json", SourceSegmentIndex: &two, Speaker: "Bob", Start: 12, End: 13, Text: "b"},
+			{ID: 3, Source: "c.json", SourceSegmentIndex: &three, Speaker: "Carol", Start: 13.5, End: 15, Text: "c"},
+		},
+		OverlapGroups: []schema.OverlapGroup{{ID: 77}},
+	}
+}
+
+func boundaryFixture() schema.Transcript {
+	one := 1
+	two := 2
+	return schema.Transcript{
+		Metadata: schema.Metadata{
+			Application: "seriatim",
+			Version:     "v-test",
+		},
+		Segments: []schema.Segment{
+			{ID: 1, Source: "a.json", SourceSegmentIndex: &one, Speaker: "Alice", Start: 1, End: 2, Text: "a", OverlapGroupID: 7},
+			{ID: 2, Source: "b.json", SourceSegmentIndex: &two, Speaker: "Bob", Start: 2, End: 3, Text: "b", OverlapGroupID: 7},
+		},
+		OverlapGroups: []schema.OverlapGroup{{ID: 7, Start: 1, End: 3}},
+	}
+}
+
+func assertSegmentIDs(t *testing.T, segments []schema.Segment, want []int) {
+	t.Helper()
+	got := make([]int, len(segments))
+	for index, segment := range segments {
+		got[index] = segment.ID
+	}
+	assertIntSlice(t, got, want)
+}
+
+func assertSegmentTexts(t *testing.T, segments []schema.Segment, want []string) {
+	t.Helper()
+	got := make([]string, len(segments))
+	for index, segment := range segments {
+		got[index] = segment.Text
+	}
+	if !equalStringSlices(got, want) {
+		t.Fatalf("segment texts = %v, want %v", got, want)
+	}
+}
+
+func assertIntSlice(t *testing.T, got []int, want []int) {
+	t.Helper()
+	if len(got) != len(want) {
+		t.Fatalf("slice length = %d, want %d", len(got), len(want))
+	}
+	for index := range got {
+		if got[index] != want[index] {
+			t.Fatalf("slice[%d] = %d, want %d (full got=%v, want=%v)", index, got[index], want[index], got, want)
+		}
+	}
+}
+
+func assertIntMap(t *testing.T, got map[int]int, want map[int]int) {
+	t.Helper()
+	if len(got) != len(want) {
+		t.Fatalf("map length = %d, want %d", len(got), len(want))
+	}
+	for key, wantValue := range want {
+		gotValue, exists := got[key]
+		if !exists {
+			t.Fatalf("missing map key %d", key)
+		}
+		if gotValue != wantValue {
+			t.Fatalf("map[%d] = %d, want %d", key, gotValue, wantValue)
+		}
+	}
+}
+
+func equalStringSlices(got []string, want []string) bool {
+	if len(got) != len(want) {
+		return false
+	}
+	for index := range got {
+		if got[index] != want[index] {
+			return false
+		}
+	}
+	return true
+}

internal/trim/artifact.go

@@ -0,0 +1,396 @@
+package trim
+
+import (
+	"encoding/json"
+	"fmt"
+
+	"gitea.maximumdirect.net/eric/seriatim/schema"
+)
+
+const (
+	SchemaMinimal      = "seriatim-minimal"
+	SchemaIntermediate = "seriatim-intermediate"
+	SchemaFull         = "seriatim-full"
+)
+
+// Artifact stores a parsed seriatim output artifact of one supported schema.
+type Artifact struct {
+	Schema       string
+	Full         *schema.Transcript
+	Intermediate *schema.IntermediateTranscript
+	Minimal      *schema.MinimalTranscript
+}
+
+// ApplyArtifactResult contains trimmed artifact output and ID mapping metadata.
+type ApplyArtifactResult struct {
+	Artifact                Artifact
+	OldToNewID              map[int]int
+	RemovedIDs              []int
+	OverlapGroupsRecomputed bool
+}
+
+// ParseArtifactJSON parses and validates a serialized seriatim output artifact.
+func ParseArtifactJSON(data []byte) (Artifact, error) {
+	var decoded any
+	if err := json.Unmarshal(data, &decoded); err != nil {
+		return Artifact{}, fmt.Errorf("input JSON is malformed: %w", err)
+	}
+
+	var full schema.Transcript
+	if err := json.Unmarshal(data, &full); err == nil {
+		if err := schema.ValidateTranscript(full); err == nil {
+			return Artifact{
+				Schema: SchemaFull,
+				Full:   &full,
+			}, nil
+		}
+	}
+
+	var intermediate schema.IntermediateTranscript
+	if err := json.Unmarshal(data, &intermediate); err == nil {
+		if err := schema.ValidateIntermediateTranscript(intermediate); err == nil {
+			return Artifact{
+				Schema:       SchemaIntermediate,
+				Intermediate: &intermediate,
+			}, nil
+		}
+	}
+
+	var minimal schema.MinimalTranscript
+	if err := json.Unmarshal(data, &minimal); err == nil {
+		if err := schema.ValidateMinimalTranscript(minimal); err == nil {
+			return Artifact{
+				Schema:  SchemaMinimal,
+				Minimal: &minimal,
+			}, nil
+		}
+	}
+
+	return Artifact{}, fmt.Errorf("input JSON is not a valid seriatim output artifact")
+}
+
+// ValidateArtifact validates an artifact against its declared schema.
+func ValidateArtifact(artifact Artifact) error {
+	switch artifact.Schema {
+	case SchemaFull:
+		if artifact.Full == nil {
+			return fmt.Errorf("full artifact payload is missing")
+		}
+		return schema.ValidateTranscript(*artifact.Full)
+	case SchemaIntermediate:
+		if artifact.Intermediate == nil {
+			return fmt.Errorf("intermediate artifact payload is missing")
+		}
+		return schema.ValidateIntermediateTranscript(*artifact.Intermediate)
+	case SchemaMinimal:
+		if artifact.Minimal == nil {
+			return fmt.Errorf("minimal artifact payload is missing")
+		}
+		return schema.ValidateMinimalTranscript(*artifact.Minimal)
+	default:
+		return fmt.Errorf("unsupported artifact schema %q", artifact.Schema)
+	}
+}
+
+// Value returns the artifact value for JSON serialization.
+func (artifact Artifact) Value() any {
+	switch artifact.Schema {
+	case SchemaFull:
+		if artifact.Full == nil {
+			return schema.Transcript{}
+		}
+		return *artifact.Full
+	case SchemaIntermediate:
+		if artifact.Intermediate == nil {
+			return schema.IntermediateTranscript{}
+		}
+		return *artifact.Intermediate
+	case SchemaMinimal:
+		if artifact.Minimal == nil {
+			return schema.MinimalTranscript{}
+		}
+		return *artifact.Minimal
+	default:
+		return nil
+	}
+}
+
+// SegmentCount returns the number of segments in the artifact.
+func (artifact Artifact) SegmentCount() int {
+	switch artifact.Schema {
+	case SchemaFull:
+		if artifact.Full == nil {
+			return 0
+		}
+		return len(artifact.Full.Segments)
+	case SchemaIntermediate:
+		if artifact.Intermediate == nil {
+			return 0
+		}
+		return len(artifact.Intermediate.Segments)
+	case SchemaMinimal:
+		if artifact.Minimal == nil {
+			return 0
+		}
+		return len(artifact.Minimal.Segments)
+	default:
+		return 0
+	}
+}
+
+// Application returns artifact metadata application name.
+func (artifact Artifact) Application() string {
+	switch artifact.Schema {
+	case SchemaFull:
+		if artifact.Full == nil {
+			return ""
+		}
+		return artifact.Full.Metadata.Application
+	case SchemaIntermediate:
+		if artifact.Intermediate == nil {
+			return ""
+		}
+		return artifact.Intermediate.Metadata.Application
+	case SchemaMinimal:
+		if artifact.Minimal == nil {
+			return ""
+		}
+		return artifact.Minimal.Metadata.Application
+	default:
+		return ""
+	}
+}
+
+// Version returns artifact metadata version.
+func (artifact Artifact) Version() string {
+	switch artifact.Schema {
+	case SchemaFull:
+		if artifact.Full == nil {
+			return ""
+		}
+		return artifact.Full.Metadata.Version
+	case SchemaIntermediate:
+		if artifact.Intermediate == nil {
+			return ""
+		}
+		return artifact.Intermediate.Metadata.Version
+	case SchemaMinimal:
+		if artifact.Minimal == nil {
+			return ""
+		}
+		return artifact.Minimal.Metadata.Version
+	default:
+		return ""
+	}
+}
+
+// ApplyArtifact trims a parsed artifact while preserving its input schema.
+func ApplyArtifact(input Artifact, opts Options) (ApplyArtifactResult, error) {
+	switch input.Schema {
+	case SchemaFull:
+		if input.Full == nil {
+			return ApplyArtifactResult{}, fmt.Errorf("full artifact payload is missing")
+		}
+		result, err := Apply(*input.Full, opts)
+		if err != nil {
+			return ApplyArtifactResult{}, err
+		}
+		out := result.Transcript
+		return ApplyArtifactResult{
+			Artifact: Artifact{
+				Schema: SchemaFull,
+				Full:   &out,
+			},
+			OldToNewID:              result.OldToNewID,
+			RemovedIDs:              result.RemovedIDs,
+			OverlapGroupsRecomputed: true,
+		}, nil
+	case SchemaIntermediate:
+		if input.Intermediate == nil {
+			return ApplyArtifactResult{}, fmt.Errorf("intermediate artifact payload is missing")
+		}
+		result, err := ApplyIntermediate(*input.Intermediate, opts)
+		if err != nil {
+			return ApplyArtifactResult{}, err
+		}
+		out := result.Transcript
+		return ApplyArtifactResult{
+			Artifact: Artifact{
+				Schema:       SchemaIntermediate,
+				Intermediate: &out,
+			},
+			OldToNewID:              result.OldToNewID,
+			RemovedIDs:              result.RemovedIDs,
+			OverlapGroupsRecomputed: false,
+		}, nil
+	case SchemaMinimal:
+		if input.Minimal == nil {
+			return ApplyArtifactResult{}, fmt.Errorf("minimal artifact payload is missing")
+		}
+		result, err := ApplyMinimal(*input.Minimal, opts)
+		if err != nil {
+			return ApplyArtifactResult{}, err
+		}
+		out := result.Transcript
+		return ApplyArtifactResult{
+			Artifact: Artifact{
+				Schema:  SchemaMinimal,
+				Minimal: &out,
+			},
+			OldToNewID:              result.OldToNewID,
+			RemovedIDs:              result.RemovedIDs,
+			OverlapGroupsRecomputed: false,
+		}, nil
+	default:
+		return ApplyArtifactResult{}, fmt.Errorf("unsupported artifact schema %q", input.Schema)
+	}
+}
+
+// ConvertArtifact converts a parsed artifact to another supported output schema.
+func ConvertArtifact(input Artifact, outputSchema string) (Artifact, error) {
+	if outputSchema == "" || outputSchema == input.Schema {
+		return input, nil
+	}
+
+	switch input.Schema {
+	case SchemaFull:
+		if input.Full == nil {
+			return Artifact{}, fmt.Errorf("full artifact payload is missing")
+		}
+		switch outputSchema {
+		case SchemaIntermediate:
+			out := intermediateFromFull(*input.Full)
+			return Artifact{
+				Schema:       SchemaIntermediate,
+				Intermediate: &out,
+			}, nil
+		case SchemaMinimal:
+			out := minimalFromFull(*input.Full)
+			return Artifact{
+				Schema:  SchemaMinimal,
+				Minimal: &out,
+			}, nil
+		default:
+			return Artifact{}, fmt.Errorf("unsupported output schema %q", outputSchema)
+		}
+	case SchemaIntermediate:
+		if input.Intermediate == nil {
+			return Artifact{}, fmt.Errorf("intermediate artifact payload is missing")
+		}
+		switch outputSchema {
+		case SchemaMinimal:
+			out := minimalFromIntermediate(*input.Intermediate)
+			return Artifact{
+				Schema:  SchemaMinimal,
+				Minimal: &out,
+			}, nil
+		case SchemaFull:
+			return Artifact{}, fmt.Errorf("cannot emit %q from %q input artifact", SchemaFull, SchemaIntermediate)
+		default:
+			return Artifact{}, fmt.Errorf("unsupported output schema %q", outputSchema)
+		}
+	case SchemaMinimal:
+		if input.Minimal == nil {
+			return Artifact{}, fmt.Errorf("minimal artifact payload is missing")
+		}
+		switch outputSchema {
+		case SchemaIntermediate:
+			out := intermediateFromMinimal(*input.Minimal)
+			return Artifact{
+				Schema:       SchemaIntermediate,
+				Intermediate: &out,
+			}, nil
+		case SchemaFull:
+			return Artifact{}, fmt.Errorf("cannot emit %q from %q input artifact", SchemaFull, SchemaMinimal)
+		default:
+			return Artifact{}, fmt.Errorf("unsupported output schema %q", outputSchema)
+		}
+	default:
+		return Artifact{}, fmt.Errorf("unsupported input schema %q", input.Schema)
+	}
+}
+
+func intermediateFromFull(input schema.Transcript) schema.IntermediateTranscript {
+	segments := make([]schema.IntermediateSegment, len(input.Segments))
+	for index, segment := range input.Segments {
+		segments[index] = schema.IntermediateSegment{
+			ID:         segment.ID,
+			Start:      segment.Start,
+			End:        segment.End,
+			Speaker:    segment.Speaker,
+			Text:       segment.Text,
+			Categories: append([]string(nil), segment.Categories...),
+		}
+	}
+	return schema.IntermediateTranscript{
+		Metadata: schema.IntermediateMetadata{
+			Application:  input.Metadata.Application,
+			Version:      input.Metadata.Version,
+			OutputSchema: SchemaIntermediate,
+		},
+		Segments: segments,
+	}
+}
+
+func minimalFromFull(input schema.Transcript) schema.MinimalTranscript {
+	segments := make([]schema.MinimalSegment, len(input.Segments))
+	for index, segment := range input.Segments {
+		segments[index] = schema.MinimalSegment{
+			ID:      segment.ID,
+			Start:   segment.Start,
+			End:     segment.End,
+			Speaker: segment.Speaker,
+			Text:    segment.Text,
+		}
+	}
+	return schema.MinimalTranscript{
+		Metadata: schema.MinimalMetadata{
+			Application:  input.Metadata.Application,
+			Version:      input.Metadata.Version,
+			OutputSchema: SchemaMinimal,
+		},
+		Segments: segments,
+	}
+}
+
+func minimalFromIntermediate(input schema.IntermediateTranscript) schema.MinimalTranscript {
+	segments := make([]schema.MinimalSegment, len(input.Segments))
+	for index, segment := range input.Segments {
+		segments[index] = schema.MinimalSegment{
+			ID:      segment.ID,
+			Start:   segment.Start,
+			End:     segment.End,
+			Speaker: segment.Speaker,
+			Text:    segment.Text,
+		}
+	}
+	return schema.MinimalTranscript{
+		Metadata: schema.MinimalMetadata{
+			Application:  input.Metadata.Application,
+			Version:      input.Metadata.Version,
+			OutputSchema: SchemaMinimal,
+		},
+		Segments: segments,
+	}
+}
+
+func intermediateFromMinimal(input schema.MinimalTranscript) schema.IntermediateTranscript {
+	segments := make([]schema.IntermediateSegment, len(input.Segments))
+	for index, segment := range input.Segments {
+		segments[index] = schema.IntermediateSegment{
+			ID:      segment.ID,
+			Start:   segment.Start,
+			End:     segment.End,
+			Speaker: segment.Speaker,
+			Text:    segment.Text,
+		}
+	}
+	return schema.IntermediateTranscript{
+		Metadata: schema.IntermediateMetadata{
+			Application:  input.Metadata.Application,
+			Version:      input.Metadata.Version,
+			OutputSchema: SchemaIntermediate,
+		},
+		Segments: segments,
+	}
+}

internal/trim/artifact_test.go

@@ -0,0 +1,138 @@
+package trim
+
+import (
+	"encoding/json"
+	"strings"
+	"testing"
+
+	"gitea.maximumdirect.net/eric/seriatim/schema"
+)
+
+func TestParseArtifactJSONRejectsMalformedJSON(t *testing.T) {
+	_, err := ParseArtifactJSON([]byte(`{"metadata":`))
+	if err == nil {
+		t.Fatal("expected malformed JSON error")
+	}
+	if !strings.Contains(err.Error(), "input JSON is malformed") {
+		t.Fatalf("unexpected error: %v", err)
+	}
+}
+
+func TestParseArtifactJSONRejectsDuplicateSegmentIDs(t *testing.T) {
+	first := 10
+	second := 20
+	value := schema.Transcript{
+		Metadata: schema.Metadata{
+			Application: "seriatim",
+			Version:     "v-test",
+		},
+		Segments: []schema.Segment{
+			{ID: 1, Source: "a.json", SourceSegmentIndex: &first, Speaker: "A", Start: 1, End: 2, Text: "one"},
+			{ID: 1, Source: "a.json", SourceSegmentIndex: &second, Speaker: "B", Start: 2, End: 3, Text: "two"},
+		},
+		OverlapGroups: []schema.OverlapGroup{},
+	}
+	data := mustMarshalJSON(t, value)
+
+	_, err := ParseArtifactJSON(data)
+	if err == nil {
+		t.Fatal("expected invalid artifact error")
+	}
+	if !strings.Contains(err.Error(), "not a valid seriatim output artifact") {
+		t.Fatalf("unexpected error: %v", err)
+	}
+}
+
+func TestParseArtifactJSONRejectsNonSequentialSegmentIDs(t *testing.T) {
+	first := 10
+	second := 20
+	value := schema.Transcript{
+		Metadata: schema.Metadata{
+			Application: "seriatim",
+			Version:     "v-test",
+		},
+		Segments: []schema.Segment{
+			{ID: 1, Source: "a.json", SourceSegmentIndex: &first, Speaker: "A", Start: 1, End: 2, Text: "one"},
+			{ID: 3, Source: "a.json", SourceSegmentIndex: &second, Speaker: "B", Start: 2, End: 3, Text: "two"},
+		},
+		OverlapGroups: []schema.OverlapGroup{},
+	}
+	data := mustMarshalJSON(t, value)
+
+	_, err := ParseArtifactJSON(data)
+	if err == nil {
+		t.Fatal("expected invalid artifact error")
+	}
+	if !strings.Contains(err.Error(), "not a valid seriatim output artifact") {
+		t.Fatalf("unexpected error: %v", err)
+	}
+}
+
+func TestConvertArtifactMinimalToIntermediate(t *testing.T) {
+	value := schema.MinimalTranscript{
+		Metadata: schema.MinimalMetadata{
+			Application:  "seriatim",
+			Version:      "v-test",
+			OutputSchema: SchemaMinimal,
+		},
+		Segments: []schema.MinimalSegment{
+			{ID: 1, Start: 1, End: 2, Speaker: "A", Text: "one"},
+			{ID: 2, Start: 2, End: 3, Speaker: "B", Text: "two"},
+		},
+	}
+	artifact := Artifact{
+		Schema:  SchemaMinimal,
+		Minimal: &value,
+	}
+
+	converted, err := ConvertArtifact(artifact, SchemaIntermediate)
+	if err != nil {
+		t.Fatalf("convert failed: %v", err)
+	}
+	if converted.Schema != SchemaIntermediate {
+		t.Fatalf("schema = %q, want %q", converted.Schema, SchemaIntermediate)
+	}
+	if converted.Intermediate == nil {
+		t.Fatal("expected intermediate artifact")
+	}
+	if len(converted.Intermediate.Segments) != 2 {
+		t.Fatalf("segment count = %d, want 2", len(converted.Intermediate.Segments))
+	}
+	if converted.Intermediate.Segments[0].ID != 1 || converted.Intermediate.Segments[1].ID != 2 {
+		t.Fatalf("unexpected IDs: %#v", converted.Intermediate.Segments)
+	}
+}
+
+func TestConvertArtifactMinimalToFullFails(t *testing.T) {
+	value := schema.MinimalTranscript{
+		Metadata: schema.MinimalMetadata{
+			Application:  "seriatim",
+			Version:      "v-test",
+			OutputSchema: SchemaMinimal,
+		},
+		Segments: []schema.MinimalSegment{
+			{ID: 1, Start: 1, End: 2, Speaker: "A", Text: "one"},
+		},
+	}
+	artifact := Artifact{
+		Schema:  SchemaMinimal,
+		Minimal: &value,
+	}
+
+	_, err := ConvertArtifact(artifact, SchemaFull)
+	if err == nil {
+		t.Fatal("expected conversion error")
+	}
+	if !strings.Contains(err.Error(), "cannot emit") {
+		t.Fatalf("unexpected error: %v", err)
+	}
+}
+
+func mustMarshalJSON(t *testing.T, value any) []byte {
+	t.Helper()
+	data, err := json.Marshal(value)
+	if err != nil {
+		t.Fatalf("marshal: %v", err)
+	}
+	return data
+}

internal/trim/selector.go

@@ -0,0 +1,156 @@
+package trim
+
+import (
+	"fmt"
+	"regexp"
+	"sort"
+	"strconv"
+	"strings"
+)
+
+var selectorElementPattern = regexp.MustCompile(`^([+-]?\d+)(?:\s*-\s*([+-]?\d+))?$`)
+
+// Selector represents a normalized union of segment IDs.
+type Selector struct {
+	ranges []idRange
+}
+
+type idRange struct {
+	start int
+	end   int
+}
+
+// ParseSelector parses an inline segment selector expression.
+func ParseSelector(input string) (Selector, error) {
+	if strings.TrimSpace(input) == "" {
+		return Selector{}, fmt.Errorf("selector cannot be empty")
+	}
+
+	parts := strings.Split(input, ",")
+	ranges := make([]idRange, 0, len(parts))
+	for index, raw := range parts {
+		element := strings.TrimSpace(raw)
+		if element == "" {
+			return Selector{}, fmt.Errorf("selector element %d cannot be empty", index+1)
+		}
+
+		rangeValue, err := parseElement(element)
+		if err != nil {
+			return Selector{}, fmt.Errorf("selector element %d %q: %w", index+1, element, err)
+		}
+		ranges = append(ranges, rangeValue)
+	}
+
+	normalized := normalizeRanges(ranges)
+	if len(normalized) == 0 {
+		return Selector{}, fmt.Errorf("selector cannot be empty")
+	}
+	return Selector{ranges: normalized}, nil
+}
+
+// Contains returns true when id is included by this selector.
+func (s Selector) Contains(id int) bool {
+	if id <= 0 {
+		return false
+	}
+	index := sort.Search(len(s.ranges), func(i int) bool {
+		return s.ranges[i].end >= id
+	})
+	if index == len(s.ranges) {
+		return false
+	}
+	rangeValue := s.ranges[index]
+	return id >= rangeValue.start && id <= rangeValue.end
+}
+
+// IDs returns a deterministic ascending list of unique segment IDs.
+func (s Selector) IDs() []int {
+	total := 0
+	for _, rangeValue := range s.ranges {
+		total += rangeValue.end - rangeValue.start + 1
+	}
+
+	ids := make([]int, 0, total)
+	for _, rangeValue := range s.ranges {
+		for id := rangeValue.start; id <= rangeValue.end; id++ {
+			ids = append(ids, id)
+		}
+	}
+	return ids
+}
+
+func parseElement(element string) (idRange, error) {
+	matches := selectorElementPattern.FindStringSubmatch(element)
+	if matches == nil {
+		return idRange{}, fmt.Errorf("malformed element")
+	}
+
+	start, err := parseID(matches[1])
+	if err != nil {
+		return idRange{}, err
+	}
+
+	if matches[2] == "" {
+		return idRange{start: start, end: start}, nil
+	}
+
+	end, err := parseID(matches[2])
+	if err != nil {
+		return idRange{}, fmt.Errorf("invalid range end: %w", err)
+	}
+	if start > end {
+		return idRange{}, fmt.Errorf("descending range %d-%d is invalid", start, end)
+	}
+	return idRange{start: start, end: end}, nil
+}
+
+func parseID(value string) (int, error) {
+	value = strings.TrimSpace(value)
+	if value == "" {
+		return 0, fmt.Errorf("missing segment ID")
+	}
+
+	id, err := strconv.Atoi(value)
+	if err != nil {
+		return 0, fmt.Errorf("segment ID must be an integer")
+	}
+	if id <= 0 {
+		return 0, fmt.Errorf("segment ID must be positive")
+	}
+	return id, nil
+}
+
+func normalizeRanges(in []idRange) []idRange {
+	if len(in) == 0 {
+		return nil
+	}
+
+	sorted := make([]idRange, len(in))
+	copy(sorted, in)
+	sort.Slice(sorted, func(i, j int) bool {
+		if sorted[i].start == sorted[j].start {
+			return sorted[i].end < sorted[j].end
+		}
+		return sorted[i].start < sorted[j].start
+	})
+
+	merged := make([]idRange, 0, len(sorted))
+	for _, next := range sorted {
+		if len(merged) == 0 {
+			merged = append(merged, next)
+			continue
+		}
+
+		last := &merged[len(merged)-1]
+		if next.start <= last.end+1 {
+			if next.end > last.end {
+				last.end = next.end
+			}
+			continue
+		}
+
+		merged = append(merged, next)
+	}
+
+	return merged
+}

internal/trim/selector_test.go

@@ -0,0 +1,127 @@
+package trim
+
+import (
+	"strings"
+	"testing"
+)
+
+func TestParseSelectorSingleID(t *testing.T) {
+	selector, err := ParseSelector("1")
+	if err != nil {
+		t.Fatalf("parse failed: %v", err)
+	}
+	assertIDs(t, selector, []int{1})
+	assertContains(t, selector, map[int]bool{1: true, 2: false, 0: false, -1: false})
+}
+
+func TestParseSelectorInclusiveRange(t *testing.T) {
+	selector, err := ParseSelector("1-3")
+	if err != nil {
+		t.Fatalf("parse failed: %v", err)
+	}
+	assertIDs(t, selector, []int{1, 2, 3})
+}
+
+func TestParseSelectorCommaSeparatedCombination(t *testing.T) {
+	selector, err := ParseSelector("1-3,8,10-12")
+	if err != nil {
+		t.Fatalf("parse failed: %v", err)
+	}
+	assertIDs(t, selector, []int{1, 2, 3, 8, 10, 11, 12})
+}
+
+func TestParseSelectorWhitespaceTolerance(t *testing.T) {
+	selector, err := ParseSelector(" 1 - 3 ,  8 , 10 - 12 ")
+	if err != nil {
+		t.Fatalf("parse failed: %v", err)
+	}
+	assertIDs(t, selector, []int{1, 2, 3, 8, 10, 11, 12})
+}
+
+func TestParseSelectorDuplicatesAndOverlapsNormalizeUnion(t *testing.T) {
+	selector, err := ParseSelector("1-4,2,4,3-6,6")
+	if err != nil {
+		t.Fatalf("parse failed: %v", err)
+	}
+	assertIDs(t, selector, []int{1, 2, 3, 4, 5, 6})
+	assertContains(t, selector, map[int]bool{1: true, 5: true, 6: true, 7: false})
+}
+
+func TestParseSelectorDeterministicNormalizedOutput(t *testing.T) {
+	left, err := ParseSelector("8,1-3,2,10-12")
+	if err != nil {
+		t.Fatalf("parse left failed: %v", err)
+	}
+	right, err := ParseSelector("10-12,3,2,1,8")
+	if err != nil {
+		t.Fatalf("parse right failed: %v", err)
+	}
+
+	leftIDs := left.IDs()
+	rightIDs := right.IDs()
+	if !equalInts(leftIDs, rightIDs) {
+		t.Fatalf("normalized IDs mismatch: %v vs %v", leftIDs, rightIDs)
+	}
+}
+
+func TestParseSelectorFailures(t *testing.T) {
+	tests := []struct {
+		name      string
+		selector  string
+		wantError string
+	}{
+		{name: "empty", selector: "", wantError: "cannot be empty"},
+		{name: "whitespace only", selector: "   ", wantError: "cannot be empty"},
+		{name: "zero", selector: "0", wantError: "must be positive"},
+		{name: "negative", selector: "-1", wantError: "must be positive"},
+		{name: "range includes zero", selector: "0-2", wantError: "must be positive"},
+		{name: "descending range", selector: "10-1", wantError: "descending range"},
+		{name: "empty element", selector: "1,,2", wantError: "cannot be empty"},
+		{name: "trailing comma", selector: "1,", wantError: "cannot be empty"},
+		{name: "malformed alpha", selector: "abc", wantError: "malformed element"},
+		{name: "malformed range", selector: "1-2-3", wantError: "malformed element"},
+		{name: "missing end", selector: "1-", wantError: "malformed element"},
+		{name: "missing start", selector: "-2", wantError: "must be positive"},
+	}
+
+	for _, test := range tests {
+		t.Run(test.name, func(t *testing.T) {
+			_, err := ParseSelector(test.selector)
+			if err == nil {
+				t.Fatalf("expected error for %q", test.selector)
+			}
+			if !strings.Contains(err.Error(), test.wantError) {
+				t.Fatalf("error = %q, want substring %q", err.Error(), test.wantError)
+			}
+		})
+	}
+}
+
+func assertIDs(t *testing.T, selector Selector, want []int) {
+	t.Helper()
+	got := selector.IDs()
+	if !equalInts(got, want) {
+		t.Fatalf("IDs = %v, want %v", got, want)
+	}
+}
+
+func assertContains(t *testing.T, selector Selector, checks map[int]bool) {
+	t.Helper()
+	for id, want := range checks {
+		if got := selector.Contains(id); got != want {
+			t.Fatalf("Contains(%d) = %t, want %t", id, got, want)
+		}
+	}
+}
+
+func equalInts(left []int, right []int) bool {
+	if len(left) != len(right) {
+		return false
+	}
+	for index := range left {
+		if left[index] != right[index] {
+			return false
+		}
+	}
+	return true
+}

schema/full-output.schema.json

@@ -1,6 +1,6 @@
 {
   "$schema": "https://json-schema.org/draft/2020-12/schema",
-  "$id": "https://gitea.maximumdirect.net/eric/seriatim/schema/output.schema.json",
+  "$id": "https://gitea.maximumdirect.net/eric/seriatim/schema/full-output.schema.json",
   "title": "seriatim full output transcript",
   "type": "object",
   "additionalProperties": false,

schema/intermediate-output.schema.json

@@ -1,7 +1,7 @@
 {
   "$schema": "https://json-schema.org/draft/2020-12/schema",
-  "$id": "https://gitea.maximumdirect.net/eric/seriatim/schema/default-output.schema.json",
-  "title": "seriatim default output transcript",
+  "$id": "https://gitea.maximumdirect.net/eric/seriatim/schema/intermediate-output.schema.json",
+  "title": "seriatim intermediate output transcript",
   "type": "object",
   "additionalProperties": false,
   "required": ["metadata", "segments"],
@@ -13,7 +13,7 @@
       "properties": {
         "application": { "type": "string" },
         "version": { "type": "string" },
-        "output_schema": { "type": "string", "const": "default" }
+        "output_schema": { "type": "string", "const": "seriatim-intermediate" }
       }
     },
     "segments": {

schema/minimal-output.schema.json

@@ -13,7 +13,7 @@
       "properties": {
         "application": { "type": "string" },
         "version": { "type": "string" },
-        "output_schema": { "type": "string", "const": "minimal" }
+        "output_schema": { "type": "string", "const": "seriatim-minimal" }
       }
     },
     "segments": {

schema/output.go

@@ -14,8 +14,8 @@ import (
 var schemaFS embed.FS
 
 const (
-	outputSchemaPath        = "output.schema.json"
-	defaultOutputSchemaPath = "default-output.schema.json"
+	fullOutputSchemaPath         = "full-output.schema.json"
+	intermediateOutputSchemaPath = "intermediate-output.schema.json"
 	minimalOutputSchemaPath      = "minimal-output.schema.json"
 )
 
@@ -32,10 +32,10 @@ type Transcript struct {
 	OverlapGroups []OverlapGroup `json:"overlap_groups"`
 }
 
-// DefaultTranscript is seriatim's default public JSON output contract.
-type DefaultTranscript struct {
-	Metadata DefaultMetadata  `json:"metadata"`
-	Segments []DefaultSegment `json:"segments"`
+// IntermediateTranscript is seriatim's intermediate public JSON output contract.
+type IntermediateTranscript struct {
+	Metadata IntermediateMetadata  `json:"metadata"`
+	Segments []IntermediateSegment `json:"segments"`
 }
 
 // MinimalTranscript is seriatim's compact public JSON output contract.
@@ -55,8 +55,8 @@ type Metadata struct {
 	OutputModules         []string `json:"output_modules"`
 }
 
-// DefaultMetadata records default artifact identity.
-type DefaultMetadata struct {
+// IntermediateMetadata records intermediate artifact identity.
+type IntermediateMetadata struct {
 	Application  string `json:"application"`
 	Version      string `json:"version"`
 	OutputSchema string `json:"output_schema"`
@@ -84,9 +84,9 @@ type Segment struct {
 	OverlapGroupID     int      `json:"overlap_group_id,omitempty"`
 }
 
-// DefaultSegment is the compact public transcript segment shape with
+// IntermediateSegment is the compact public transcript segment shape with
 // categories.
-type DefaultSegment struct {
+type IntermediateSegment struct {
 	ID         int      `json:"id"`
 	Start      float64  `json:"start"`
 	End        float64  `json:"end"`
@@ -115,7 +115,7 @@ type OverlapGroup struct {
 	Resolution string   `json:"resolution"`
 }
 
-// ValidateTranscript validates a typed transcript against the public JSON
+// ValidateTranscript validates a full transcript against the public JSON
 // schema and seriatim-specific semantic rules.
 func ValidateTranscript(transcript Transcript) error {
 	if err := validateSemantics(transcript); err != nil {
@@ -129,18 +129,18 @@ func ValidateTranscript(transcript Transcript) error {
 	return ValidateJSON(data)
 }
 
-// ValidateDefaultTranscript validates the default transcript against the
-// default JSON schema and seriatim-specific semantic rules.
-func ValidateDefaultTranscript(transcript DefaultTranscript) error {
-	if err := validateDefaultSemantics(transcript); err != nil {
+// ValidateIntermediateTranscript validates the intermediate transcript against
+// the intermediate JSON schema and seriatim-specific semantic rules.
+func ValidateIntermediateTranscript(transcript IntermediateTranscript) error {
+	if err := validateIntermediateSemantics(transcript); err != nil {
 		return err
 	}
 
 	data, err := json.Marshal(transcript)
 	if err != nil {
-		return fmt.Errorf("marshal default transcript for schema validation: %w", err)
+		return fmt.Errorf("marshal intermediate transcript for schema validation: %w", err)
 	}
-	return ValidateDefaultJSON(data)
+	return ValidateIntermediateJSON(data)
 }
 
 // ValidateMinimalTranscript validates a minimal transcript against the minimal
@@ -159,13 +159,13 @@ func ValidateMinimalTranscript(transcript MinimalTranscript) error {
 
 // ValidateJSON validates serialized output JSON against the public schema.
 func ValidateJSON(data []byte) error {
-	return validateJSONWithSchema(data, outputSchemaPath)
+	return validateJSONWithSchema(data, fullOutputSchemaPath)
 }
 
-// ValidateDefaultJSON validates serialized default output JSON against the
-// default public schema.
-func ValidateDefaultJSON(data []byte) error {
-	return validateJSONWithSchema(data, defaultOutputSchemaPath)
+// ValidateIntermediateJSON validates serialized intermediate output JSON
+// against the intermediate public schema.
+func ValidateIntermediateJSON(data []byte) error {
+	return validateJSONWithSchema(data, intermediateOutputSchemaPath)
 }
 
 // ValidateMinimalJSON validates serialized minimal output JSON against the
@@ -245,7 +245,7 @@ func validateSemantics(transcript Transcript) error {
 	return nil
 }
 
-func validateDefaultSemantics(transcript DefaultTranscript) error {
+func validateIntermediateSemantics(transcript IntermediateTranscript) error {
 	for index, segment := range transcript.Segments {
 		wantID := index + 1
 		if segment.ID != wantID {

schema/output_test.go

@@ -21,11 +21,11 @@ func TestValidateMinimalTranscriptAcceptsValidTranscript(t *testing.T) {
 	}
 }
 
-func TestValidateDefaultTranscriptAcceptsValidTranscript(t *testing.T) {
-	transcript := validDefaultTranscript()
+func TestValidateIntermediateTranscriptAcceptsValidTranscript(t *testing.T) {
+	transcript := validIntermediateTranscript()
 
-	if err := ValidateDefaultTranscript(transcript); err != nil {
-		t.Fatalf("validate default transcript: %v", err)
+	if err := ValidateIntermediateTranscript(transcript); err != nil {
+		t.Fatalf("validate intermediate transcript: %v", err)
 	}
 }
 
@@ -34,7 +34,7 @@ func TestValidateMinimalJSONRejectsMissingRequiredField(t *testing.T) {
 		"metadata": {
 			"application": "seriatim",
 			"version": "dev",
-			"output_schema": "minimal"
+				"output_schema": "seriatim-minimal"
 		}
 	}`))
 	assertErrorContains(t, err, "segments")
@@ -45,7 +45,7 @@ func TestValidateMinimalJSONRejectsWrongFieldType(t *testing.T) {
 		"metadata": {
 			"application": "seriatim",
 			"version": "dev",
-			"output_schema": "minimal"
+				"output_schema": "seriatim-minimal"
 		},
 		"segments": [
 			{
@@ -68,7 +68,7 @@ func TestValidateMinimalJSONRejectsUnexpectedFields(t *testing.T) {
 		{
 			name: "top-level overlap groups",
 			json: `{
-				"metadata": {"application": "seriatim", "version": "dev", "output_schema": "minimal"},
+					"metadata": {"application": "seriatim", "version": "dev", "output_schema": "seriatim-minimal"},
 				"segments": [],
 				"overlap_groups": []
 			}`,
@@ -76,28 +76,28 @@ func TestValidateMinimalJSONRejectsUnexpectedFields(t *testing.T) {
 		{
 			name: "segment source",
 			json: `{
-				"metadata": {"application": "seriatim", "version": "dev", "output_schema": "minimal"},
+					"metadata": {"application": "seriatim", "version": "dev", "output_schema": "seriatim-minimal"},
 				"segments": [{"id": 1, "source": "input.json", "start": 1, "end": 2, "speaker": "Alice", "text": "hello"}]
 			}`,
 		},
 		{
 			name: "segment categories",
 			json: `{
-				"metadata": {"application": "seriatim", "version": "dev", "output_schema": "minimal"},
+					"metadata": {"application": "seriatim", "version": "dev", "output_schema": "seriatim-minimal"},
 				"segments": [{"id": 1, "start": 1, "end": 2, "speaker": "Alice", "text": "hello", "categories": ["backchannel"]}]
 			}`,
 		},
 		{
 			name: "segment derived from",
 			json: `{
-				"metadata": {"application": "seriatim", "version": "dev", "output_schema": "minimal"},
+					"metadata": {"application": "seriatim", "version": "dev", "output_schema": "seriatim-minimal"},
 				"segments": [{"id": 1, "start": 1, "end": 2, "speaker": "Alice", "text": "hello", "derived_from": ["input.json#0"]}]
 			}`,
 		},
 		{
 			name: "segment words",
 			json: `{
-				"metadata": {"application": "seriatim", "version": "dev", "output_schema": "minimal"},
+					"metadata": {"application": "seriatim", "version": "dev", "output_schema": "seriatim-minimal"},
 				"segments": [{"id": 1, "start": 1, "end": 2, "speaker": "Alice", "text": "hello", "words": []}]
 			}`,
 		},
@@ -111,23 +111,23 @@ func TestValidateMinimalJSONRejectsUnexpectedFields(t *testing.T) {
 	}
 }
 
-func TestValidateDefaultJSONRejectsMissingRequiredField(t *testing.T) {
-	err := ValidateDefaultJSON([]byte(`{
+func TestValidateIntermediateJSONRejectsMissingRequiredField(t *testing.T) {
+	err := ValidateIntermediateJSON([]byte(`{
 			"metadata": {
 				"application": "seriatim",
 				"version": "dev",
-			"output_schema": "default"
+				"output_schema": "seriatim-intermediate"
 			}
 		}`))
 	assertErrorContains(t, err, "segments")
 }
 
-func TestValidateDefaultJSONRejectsWrongFieldType(t *testing.T) {
-	err := ValidateDefaultJSON([]byte(`{
+func TestValidateIntermediateJSONRejectsWrongFieldType(t *testing.T) {
+	err := ValidateIntermediateJSON([]byte(`{
 			"metadata": {
 				"application": "seriatim",
 				"version": "dev",
-			"output_schema": "default"
+				"output_schema": "seriatim-intermediate"
 			},
 		"segments": [
 			{
@@ -142,7 +142,7 @@ func TestValidateDefaultJSONRejectsWrongFieldType(t *testing.T) {
 	assertErrorContains(t, err, "id")
 }
 
-func TestValidateDefaultJSONRejectsUnexpectedFields(t *testing.T) {
+func TestValidateIntermediateJSONRejectsUnexpectedFields(t *testing.T) {
 	tests := []struct {
 		name string
 		json string
@@ -150,7 +150,7 @@ func TestValidateDefaultJSONRejectsUnexpectedFields(t *testing.T) {
 		{
 			name: "top-level overlap groups",
 			json: `{
-				"metadata": {"application": "seriatim", "version": "dev", "output_schema": "default"},
+					"metadata": {"application": "seriatim", "version": "dev", "output_schema": "seriatim-intermediate"},
 				"segments": [],
 				"overlap_groups": []
 			}`,
@@ -158,21 +158,21 @@ func TestValidateDefaultJSONRejectsUnexpectedFields(t *testing.T) {
 		{
 			name: "segment source",
 			json: `{
-				"metadata": {"application": "seriatim", "version": "dev", "output_schema": "default"},
+					"metadata": {"application": "seriatim", "version": "dev", "output_schema": "seriatim-intermediate"},
 				"segments": [{"id": 1, "source": "input.json", "start": 1, "end": 2, "speaker": "Alice", "text": "hello"}]
 			}`,
 		},
 		{
 			name: "segment derived from",
 			json: `{
-				"metadata": {"application": "seriatim", "version": "dev", "output_schema": "default"},
+					"metadata": {"application": "seriatim", "version": "dev", "output_schema": "seriatim-intermediate"},
 				"segments": [{"id": 1, "start": 1, "end": 2, "speaker": "Alice", "text": "hello", "derived_from": ["input.json#0"]}]
 			}`,
 		},
 		{
 			name: "segment words",
 			json: `{
-				"metadata": {"application": "seriatim", "version": "dev", "output_schema": "default"},
+					"metadata": {"application": "seriatim", "version": "dev", "output_schema": "seriatim-intermediate"},
 				"segments": [{"id": 1, "start": 1, "end": 2, "speaker": "Alice", "text": "hello", "words": []}]
 			}`,
 		},
@@ -180,7 +180,7 @@ func TestValidateDefaultJSONRejectsUnexpectedFields(t *testing.T) {
 
 	for _, test := range tests {
 		t.Run(test.name, func(t *testing.T) {
-			err := ValidateDefaultJSON([]byte(test.json))
+			err := ValidateIntermediateJSON([]byte(test.json))
 			assertErrorContains(t, err, "additional properties")
 		})
 	}
@@ -267,7 +267,7 @@ func TestValidateJSONRejectsUnexpectedInternalFields(t *testing.T) {
 	}
 }
 
-func TestValidateDefaultTranscriptRejectsMissingOrNonSequentialIDs(t *testing.T) {
+func TestValidateIntermediateTranscriptRejectsMissingOrNonSequentialIDs(t *testing.T) {
 	tests := []struct {
 		name string
 		ids  []int
@@ -280,10 +280,10 @@ func TestValidateDefaultTranscriptRejectsMissingOrNonSequentialIDs(t *testing.T)
 
 	for _, test := range tests {
 		t.Run(test.name, func(t *testing.T) {
-			transcript := validDefaultTranscript()
+			transcript := validIntermediateTranscript()
 			transcript.Segments = transcript.Segments[:0]
 			for index, id := range test.ids {
-				transcript.Segments = append(transcript.Segments, DefaultSegment{
+				transcript.Segments = append(transcript.Segments, IntermediateSegment{
 					ID:      id,
 					Start:   float64(index),
 					End:     float64(index) + 1,
@@ -292,18 +292,18 @@ func TestValidateDefaultTranscriptRejectsMissingOrNonSequentialIDs(t *testing.T)
 				})
 			}
 
-			err := ValidateDefaultTranscript(transcript)
+			err := ValidateIntermediateTranscript(transcript)
 			assertErrorContains(t, err, test.want)
 		})
 	}
 }
 
-func TestValidateDefaultTranscriptRejectsInvalidTiming(t *testing.T) {
-	transcript := validDefaultTranscript()
+func TestValidateIntermediateTranscriptRejectsInvalidTiming(t *testing.T) {
+	transcript := validIntermediateTranscript()
 	transcript.Segments[0].Start = 2
 	transcript.Segments[0].End = 1
 
-	err := ValidateDefaultTranscript(transcript)
+	err := ValidateIntermediateTranscript(transcript)
 	assertErrorContains(t, err, "segment 0 has end")
 }
 
@@ -411,7 +411,7 @@ func validMinimalTranscript() MinimalTranscript {
 		Metadata: MinimalMetadata{
 			Application:  "seriatim",
 			Version:      "dev",
-			OutputSchema: "minimal",
+			OutputSchema: "seriatim-minimal",
 		},
 		Segments: []MinimalSegment{
 			{
@@ -425,14 +425,14 @@ func validMinimalTranscript() MinimalTranscript {
 	}
 }
 
-func validDefaultTranscript() DefaultTranscript {
-	return DefaultTranscript{
-		Metadata: DefaultMetadata{
+func validIntermediateTranscript() IntermediateTranscript {
+	return IntermediateTranscript{
+		Metadata: IntermediateMetadata{
 			Application:  "seriatim",
 			Version:      "dev",
-			OutputSchema: "default",
+			OutputSchema: "seriatim-intermediate",
 		},
-		Segments: []DefaultSegment{
+		Segments: []IntermediateSegment{
 			{
 				ID:         1,
 				Start:      1,