diff --git a/internal/llm/openai_compatible_client.go b/internal/llm/openai_compatible_client.go
new file mode 100644
index 0000000..900912a
--- /dev/null
+++ b/internal/llm/openai_compatible_client.go
@@ -0,0 +1,181 @@
+package llm
+
+import (
+	"bytes"
+	"context"
+	"encoding/json"
+	"errors"
+	"fmt"
+	"io"
+	"net/http"
+	"net/url"
+	"strings"
+	"time"
+
+	"gitea.maximumdirect.net/eric/scriptorium/internal/domain"
+)
+
+var (
+	ErrInvalidConfig     = errors.New("invalid llm client configuration")
+	ErrInvalidRequest    = errors.New("invalid generate request")
+	ErrRequestFailed     = errors.New("llm request failed")
+	ErrUnexpectedStatus  = errors.New("llm returned non-success status")
+	ErrMalformedResponse = errors.New("malformed llm response")
+)
+
+type OpenAICompatibleConfig struct {
+	BaseURL    string
+	APIKey     string
+	Model      string
+	Timeout    time.Duration
+	HTTPClient *http.Client
+}
+
+type OpenAICompatibleClient struct {
+	baseURL      string
+	apiKey       string
+	defaultModel string
+	timeout      time.Duration
+	httpClient   *http.Client
+}
+
+func NewOpenAICompatibleClient(cfg OpenAICompatibleConfig) (*OpenAICompatibleClient, error) {
+	if strings.TrimSpace(cfg.BaseURL) == "" {
+		return nil, fmt.Errorf("%w: base URL is required", ErrInvalidConfig)
+	}
+	if _, err := url.ParseRequestURI(cfg.BaseURL); err != nil {
+		return nil, fmt.Errorf("%w: invalid base URL: %v", ErrInvalidConfig, err)
+	}
+
+	timeout := cfg.Timeout
+	if timeout <= 0 {
+		timeout = 30 * time.Second
+	}
+
+	var client *http.Client
+	if cfg.HTTPClient != nil {
+		client = cfg.HTTPClient
+		if client.Timeout == 0 {
+			client.Timeout = timeout
+		}
+	} else {
+		client = &http.Client{Timeout: timeout}
+	}
+
+	return &OpenAICompatibleClient{
+		baseURL:      strings.TrimRight(cfg.BaseURL, "/"),
+		apiKey:       cfg.APIKey,
+		defaultModel: cfg.Model,
+		timeout:      timeout,
+		httpClient:   client,
+	}, nil
+}
+
+func (c *OpenAICompatibleClient) Generate(ctx context.Context, req domain.GenerateRequest) (*domain.GenerateResponse, error) {
+	model := strings.TrimSpace(req.Target.Model)
+	if model == "" {
+		model = strings.TrimSpace(c.defaultModel)
+	}
+	if model == "" {
+		return nil, fmt.Errorf("%w: model is required", ErrInvalidRequest)
+	}
+
+	endpoint := strings.TrimSpace(req.Target.Endpoint)
+	if endpoint == "" {
+		endpoint = c.baseURL
+	}
+	endpoint = strings.TrimRight(endpoint, "/") + "/chat/completions"
+
+	wireReq := openAIChatRequest{
+		Model: model,
+	}
+
+	wireReq.Messages = make([]openAIChatMessage, 0, len(req.Prompt.Messages))
+	for _, msg := range req.Prompt.Messages {
+		wireReq.Messages = append(wireReq.Messages, openAIChatMessage{
+			Role:    msg.Role,
+			Content: msg.Content,
+		})
+	}
+
+	if req.Target.Temperature != 0 {
+		wireReq.Temperature = &req.Target.Temperature
+	}
+	if req.Target.MaxTokens != 0 {
+		wireReq.MaxTokens = &req.Target.MaxTokens
+	}
+	if req.Target.TopP != 0 {
+		wireReq.TopP = &req.Target.TopP
+	}
+
+	payload, err := json.Marshal(wireReq)
+	if err != nil {
+		return nil, fmt.Errorf("%w: failed to encode request: %v", ErrRequestFailed, err)
+	}
+
+	httpReq, err := http.NewRequestWithContext(ctx, http.MethodPost, endpoint, bytes.NewReader(payload))
+	if err != nil {
+		return nil, fmt.Errorf("%w: failed to create request: %v", ErrRequestFailed, err)
+	}
+	httpReq.Header.Set("Content-Type", "application/json")
+	if strings.TrimSpace(c.apiKey) != "" {
+		httpReq.Header.Set("Authorization", "Bearer "+c.apiKey)
+	}
+
+	httpResp, err := c.httpClient.Do(httpReq)
+	if err != nil {
+		return nil, fmt.Errorf("%w: %v", ErrRequestFailed, err)
+	}
+	defer httpResp.Body.Close()
+
+	if httpResp.StatusCode < 200 || httpResp.StatusCode >= 300 {
+		body, _ := io.ReadAll(io.LimitReader(httpResp.Body, 4096))
+		return nil, fmt.Errorf("%w: status=%d body=%q", ErrUnexpectedStatus, httpResp.StatusCode, strings.TrimSpace(string(body)))
+	}
+
+	var wireResp openAIChatResponse
+	if err := json.NewDecoder(httpResp.Body).Decode(&wireResp); err != nil {
+		return nil, fmt.Errorf("%w: failed to decode response: %v", ErrMalformedResponse, err)
+	}
+
+	if len(wireResp.Choices) == 0 {
+		return nil, fmt.Errorf("%w: no choices returned", ErrMalformedResponse)
+	}
+	content := wireResp.Choices[0].Message.Content
+	if content == "" {
+		return nil, fmt.Errorf("%w: first choice has empty message content", ErrMalformedResponse)
+	}
+
+	return &domain.GenerateResponse{
+		Content: content,
+		Usage: domain.TokenUsage{
+			PromptTokens:     wireResp.Usage.PromptTokens,
+			CompletionTokens: wireResp.Usage.CompletionTokens,
+			TotalTokens:      wireResp.Usage.TotalTokens,
+		},
+	}, nil
+}
+
+type openAIChatRequest struct {
+	Model       string              `json:"model"`
+	Messages    []openAIChatMessage `json:"messages"`
+	Temperature *float64            `json:"temperature,omitempty"`
+	MaxTokens   *int                `json:"max_tokens,omitempty"`
+	TopP        *float64            `json:"top_p,omitempty"`
+}
+
+type openAIChatMessage struct {
+	Role    string `json:"role"`
+	Content string `json:"content"`
+}
+
+type openAIChatResponse struct {
+	Choices []struct {
+		Message openAIChatMessage `json:"message"`
+	} `json:"choices"`
+	Usage struct {
+		PromptTokens     int `json:"prompt_tokens"`
+		CompletionTokens int `json:"completion_tokens"`
+		TotalTokens      int `json:"total_tokens"`
+	} `json:"usage"`
+}
diff --git a/internal/llm/openai_compatible_client_test.go b/internal/llm/openai_compatible_client_test.go
new file mode 100644
index 0000000..0aa65cb
--- /dev/null
+++ b/internal/llm/openai_compatible_client_test.go
@@ -0,0 +1,289 @@
+package llm
+
+import (
+	"context"
+	"encoding/json"
+	"errors"
+	"net/http"
+	"net/http/httptest"
+	"strings"
+	"testing"
+	"time"
+
+	"gitea.maximumdirect.net/eric/scriptorium/internal/domain"
+)
+
+func TestOpenAICompatibleClientGenerateSuccess(t *testing.T) {
+	type observedRequest struct {
+		Authorization string
+		Body          map[string]any
+	}
+	obs := &observedRequest{}
+
+	ts := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		obs.Authorization = r.Header.Get("Authorization")
+		if r.URL.Path != "/v1/chat/completions" {
+			t.Fatalf("unexpected path: %s", r.URL.Path)
+		}
+		if ct := r.Header.Get("Content-Type"); ct != "application/json" {
+			t.Fatalf("unexpected content type: %s", ct)
+		}
+
+		defer r.Body.Close()
+		if err := json.NewDecoder(r.Body).Decode(&obs.Body); err != nil {
+			t.Fatalf("failed to decode request body: %v", err)
+		}
+
+		w.Header().Set("Content-Type", "application/json")
+		_, _ = w.Write([]byte(`{
+  "choices": [{"message": {"role": "assistant", "content": "hello from model"}}],
+  "usage": {"prompt_tokens": 11, "completion_tokens": 22, "total_tokens": 33}
+}`))
+	}))
+	defer ts.Close()
+
+	client, err := NewOpenAICompatibleClient(OpenAICompatibleConfig{
+		BaseURL: ts.URL + "/v1",
+		APIKey:  "secret-key",
+		Timeout: 2 * time.Second,
+	})
+	if err != nil {
+		t.Fatalf("unexpected constructor error: %v", err)
+	}
+
+	resp, err := client.Generate(context.Background(), domain.GenerateRequest{
+		Prompt: domain.RenderedPrompt{Messages: []domain.RenderedMessage{
+			{Role: "system", Content: "You are helpful."},
+			{Role: "user", Content: "Say hello"},
+		}},
+		Target: domain.ModelTarget{
+			Model:       "gpt-test",
+			Temperature: 0.4,
+			MaxTokens:   123,
+			TopP:        0.7,
+		},
+	})
+	if err != nil {
+		t.Fatalf("expected no error, got %v", err)
+	}
+
+	if resp.Content != "hello from model" {
+		t.Fatalf("unexpected content: %q", resp.Content)
+	}
+	if resp.Usage.PromptTokens != 11 || resp.Usage.CompletionTokens != 22 || resp.Usage.TotalTokens != 33 {
+		t.Fatalf("unexpected usage: %+v", resp.Usage)
+	}
+
+	if obs.Authorization != "Bearer secret-key" {
+		t.Fatalf("unexpected Authorization header: %q", obs.Authorization)
+	}
+	if got, ok := obs.Body["model"].(string); !ok || got != "gpt-test" {
+		t.Fatalf("unexpected model payload: %#v", obs.Body["model"])
+	}
+
+	msgs, ok := obs.Body["messages"].([]any)
+	if !ok || len(msgs) != 2 {
+		t.Fatalf("unexpected messages payload: %#v", obs.Body["messages"])
+	}
+	msg0 := msgs[0].(map[string]any)
+	if msg0["role"] != "system" || msg0["content"] != "You are helpful." {
+		t.Fatalf("unexpected first message: %#v", msg0)
+	}
+	msg1 := msgs[1].(map[string]any)
+	if msg1["role"] != "user" || msg1["content"] != "Say hello" {
+		t.Fatalf("unexpected second message: %#v", msg1)
+	}
+}
+
+func TestOpenAICompatibleClientNoAuthorizationHeaderWhenNoAPIKey(t *testing.T) {
+	hadAuth := false
+	ts := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		hadAuth = r.Header.Get("Authorization") != ""
+		_, _ = w.Write([]byte(`{"choices":[{"message":{"content":"ok"}}]}`))
+	}))
+	defer ts.Close()
+
+	client, err := NewOpenAICompatibleClient(OpenAICompatibleConfig{BaseURL: ts.URL + "/v1"})
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	_, err = client.Generate(context.Background(), domain.GenerateRequest{
+		Prompt: domain.RenderedPrompt{Messages: []domain.RenderedMessage{{Role: "user", Content: "hi"}}},
+		Target: domain.ModelTarget{Model: "model"},
+	})
+	if err != nil {
+		t.Fatalf("expected no error, got %v", err)
+	}
+	if hadAuth {
+		t.Fatal("did not expect Authorization header")
+	}
+}
+
+func TestOpenAICompatibleClientModelFallbackFromConfig(t *testing.T) {
+	gotModel := ""
+	ts := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		var body map[string]any
+		_ = json.NewDecoder(r.Body).Decode(&body)
+		if m, ok := body["model"].(string); ok {
+			gotModel = m
+		}
+		_, _ = w.Write([]byte(`{"choices":[{"message":{"content":"ok"}}]}`))
+	}))
+	defer ts.Close()
+
+	client, err := NewOpenAICompatibleClient(OpenAICompatibleConfig{BaseURL: ts.URL + "/v1", Model: "default-model"})
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	_, err = client.Generate(context.Background(), domain.GenerateRequest{
+		Prompt: domain.RenderedPrompt{Messages: []domain.RenderedMessage{{Role: "user", Content: "hi"}}},
+		Target: domain.ModelTarget{},
+	})
+	if err != nil {
+		t.Fatalf("expected no error, got %v", err)
+	}
+	if gotModel != "default-model" {
+		t.Fatalf("expected default model, got %q", gotModel)
+	}
+}
+
+func TestOpenAICompatibleClientEndpointOverride(t *testing.T) {
+	defaultHit := false
+	overrideHit := false
+
+	defaultServer := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		defaultHit = true
+		_, _ = w.Write([]byte(`{"choices":[{"message":{"content":"default"}}]}`))
+	}))
+	defer defaultServer.Close()
+
+	overrideServer := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		overrideHit = true
+		if r.URL.Path != "/v1/chat/completions" {
+			t.Fatalf("unexpected path: %s", r.URL.Path)
+		}
+		_, _ = w.Write([]byte(`{"choices":[{"message":{"content":"override"}}]}`))
+	}))
+	defer overrideServer.Close()
+
+	client, err := NewOpenAICompatibleClient(OpenAICompatibleConfig{BaseURL: defaultServer.URL + "/v1", Model: "m"})
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	resp, err := client.Generate(context.Background(), domain.GenerateRequest{
+		Prompt: domain.RenderedPrompt{Messages: []domain.RenderedMessage{{Role: "user", Content: "hi"}}},
+		Target: domain.ModelTarget{Endpoint: overrideServer.URL + "/v1"},
+	})
+	if err != nil {
+		t.Fatalf("expected no error, got %v", err)
+	}
+	if resp.Content != "override" {
+		t.Fatalf("expected override response, got %q", resp.Content)
+	}
+	if defaultHit {
+		t.Fatal("default endpoint should not have been called")
+	}
+	if !overrideHit {
+		t.Fatal("override endpoint should have been called")
+	}
+}
+
+func TestOpenAICompatibleClientNon2xxError(t *testing.T) {
+	ts := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		w.WriteHeader(http.StatusBadRequest)
+		_, _ = w.Write([]byte(`{"error":"bad request payload"}`))
+	}))
+	defer ts.Close()
+
+	client, err := NewOpenAICompatibleClient(OpenAICompatibleConfig{BaseURL: ts.URL + "/v1", Model: "m"})
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	_, err = client.Generate(context.Background(), domain.GenerateRequest{
+		Prompt: domain.RenderedPrompt{Messages: []domain.RenderedMessage{{Role: "user", Content: "hi"}}},
+	})
+	if err == nil {
+		t.Fatal("expected non-2xx error")
+	}
+	if !errors.Is(err, ErrUnexpectedStatus) {
+		t.Fatalf("expected ErrUnexpectedStatus, got %v", err)
+	}
+	if !strings.Contains(err.Error(), "400") || !strings.Contains(err.Error(), "bad request payload") {
+		t.Fatalf("expected status/body details, got %v", err)
+	}
+}
+
+func TestOpenAICompatibleClientMalformedResponseInvalidJSON(t *testing.T) {
+	ts := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		_, _ = w.Write([]byte(`{not valid json`))
+	}))
+	defer ts.Close()
+
+	client, err := NewOpenAICompatibleClient(OpenAICompatibleConfig{BaseURL: ts.URL + "/v1", Model: "m"})
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	_, err = client.Generate(context.Background(), domain.GenerateRequest{
+		Prompt: domain.RenderedPrompt{Messages: []domain.RenderedMessage{{Role: "user", Content: "hi"}}},
+	})
+	if err == nil {
+		t.Fatal("expected malformed response error")
+	}
+	if !errors.Is(err, ErrMalformedResponse) {
+		t.Fatalf("expected ErrMalformedResponse, got %v", err)
+	}
+}
+
+func TestOpenAICompatibleClientMalformedResponseMissingChoices(t *testing.T) {
+	ts := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		_, _ = w.Write([]byte(`{"choices": []}`))
+	}))
+	defer ts.Close()
+
+	client, err := NewOpenAICompatibleClient(OpenAICompatibleConfig{BaseURL: ts.URL + "/v1", Model: "m"})
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	_, err = client.Generate(context.Background(), domain.GenerateRequest{
+		Prompt: domain.RenderedPrompt{Messages: []domain.RenderedMessage{{Role: "user", Content: "hi"}}},
+	})
+	if err == nil {
+		t.Fatal("expected malformed response error")
+	}
+	if !errors.Is(err, ErrMalformedResponse) {
+		t.Fatalf("expected ErrMalformedResponse, got %v", err)
+	}
+}
+
+func TestOpenAICompatibleClientTimeout(t *testing.T) {
+	ts := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		time.Sleep(250 * time.Millisecond)
+		_, _ = w.Write([]byte(`{"choices":[{"message":{"content":"ok"}}]}`))
+	}))
+	defer ts.Close()
+
+	client, err := NewOpenAICompatibleClient(OpenAICompatibleConfig{
+		BaseURL: ts.URL + "/v1",
+		Model:   "m",
+		Timeout: 50 * time.Millisecond,
+	})
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	_, err = client.Generate(context.Background(), domain.GenerateRequest{
+		Prompt: domain.RenderedPrompt{Messages: []domain.RenderedMessage{{Role: "user", Content: "hi"}}},
+	})
+	if err == nil {
+		t.Fatal("expected timeout error")
+	}
+	if !errors.Is(err, ErrRequestFailed) {
+		t.Fatalf("expected ErrRequestFailed, got %v", err)
+	}
+}