feat: implement OpenAI-compatible file upload and reference handling for DeepSeek API

2026-05-14 05:05:09 +08:00 · 2026-04-12 23:30:22 +08:00
parent 0a23c77ff7
commit caafdedb00
31 changed files with 1882 additions and 330 deletions
--- a/internal/adapter/claude/proxy_vercel_test.go
+++ b/internal/adapter/claude/proxy_vercel_test.go
@@ -34,11 +34,13 @@ func (s openAIProxyStub) ChatCompletions(w http.ResponseWriter, _ *http.Request)

 type openAIProxyCaptureStub struct {
 	seenModel string
+	seenReq   map[string]any
 }

 func (s *openAIProxyCaptureStub) ChatCompletions(w http.ResponseWriter, r *http.Request) {
 	var req map[string]any
 	_ = json.NewDecoder(r.Body).Decode(&req)
+	s.seenReq = req
 	if m, ok := req["model"].(string); ok {
 		s.seenModel = m
 	}
@@ -84,3 +86,33 @@ func TestClaudeProxyViaOpenAIPreservesClaudeMapping(t *testing.T) {
 		t.Fatalf("expected mapped proxy model deepseek-reasoner, got %q", got)
 	}
 }
+
+func TestClaudeProxyTranslatesInlineImageToOpenAIDataURL(t *testing.T) {
+	openAI := &openAIProxyCaptureStub{}
+	h := &Handler{OpenAI: openAI}
+	req := httptest.NewRequest(http.MethodPost, "/anthropic/v1/messages", strings.NewReader(`{"model":"claude-sonnet-4-5","messages":[{"role":"user","content":[{"type":"text","text":"hello"},{"type":"image","source":{"type":"base64","media_type":"image/png","data":"QUJDRA=="}}]}],"stream":false}`))
+	rec := httptest.NewRecorder()
+
+	h.Messages(rec, req)
+
+	if rec.Code != http.StatusOK {
+		t.Fatalf("unexpected status: %d body=%s", rec.Code, rec.Body.String())
+	}
+	messages, _ := openAI.seenReq["messages"].([]any)
+	if len(messages) != 1 {
+		t.Fatalf("expected one translated message, got %#v", openAI.seenReq)
+	}
+	msg, _ := messages[0].(map[string]any)
+	content, _ := msg["content"].([]any)
+	if len(content) != 2 {
+		t.Fatalf("expected translated content blocks, got %#v", msg)
+	}
+	imageBlock, _ := content[1].(map[string]any)
+	if strings.TrimSpace(asString(imageBlock["type"])) != "image_url" {
+		t.Fatalf("expected image_url block, got %#v", imageBlock)
+	}
+	imageURL, _ := imageBlock["image_url"].(map[string]any)
+	if !strings.HasPrefix(strings.TrimSpace(asString(imageURL["url"])), "data:image/png;base64,") {
+		t.Fatalf("expected translated data url, got %#v", imageBlock)
+	}
+}
--- a/internal/adapter/gemini/handler_test.go
+++ b/internal/adapter/gemini/handler_test.go
@@ -82,11 +82,17 @@ func (s geminiOpenAIErrorStub) ChatCompletions(w http.ResponseWriter, _ *http.Re
 }

 type geminiOpenAISuccessStub struct {
-	stream bool
-	body   string
+	stream  bool
+	body    string
+	seenReq map[string]any
 }

-func (s geminiOpenAISuccessStub) ChatCompletions(w http.ResponseWriter, _ *http.Request) {
+func (s *geminiOpenAISuccessStub) ChatCompletions(w http.ResponseWriter, r *http.Request) {
+	if r != nil {
+		var req map[string]any
+		_ = json.NewDecoder(r.Body).Decode(&req)
+		s.seenReq = req
+	}
 	if s.stream {
 		w.Header().Set("Content-Type", "text/event-stream")
 		w.WriteHeader(http.StatusOK)
@@ -144,7 +150,7 @@ func TestGeminiRoutesRegistered(t *testing.T) {
 func TestGenerateContentReturnsFunctionCallParts(t *testing.T) {
 	h := &Handler{
 		Store: testGeminiConfig{},
-		OpenAI: geminiOpenAISuccessStub{
+		OpenAI: &geminiOpenAISuccessStub{
 			body: `{"id":"chatcmpl-1","object":"chat.completion","choices":[{"index":0,"message":{"role":"assistant","tool_calls":[{"id":"call_1","type":"function","function":{"name":"eval_javascript","arguments":"{\"code\":\"1+1\"}"}}]},"finish_reason":"tool_calls"}]}`,
 		},
 	}
@@ -184,7 +190,7 @@ func TestGenerateContentReturnsFunctionCallParts(t *testing.T) {
 }

 func TestGenerateContentMixedToolSnippetAlsoTriggersFunctionCall(t *testing.T) {
-	h := &Handler{Store: testGeminiConfig{}, OpenAI: geminiOpenAISuccessStub{}}
+	h := &Handler{Store: testGeminiConfig{}, OpenAI: &geminiOpenAISuccessStub{}}
 	r := chi.NewRouter()
 	RegisterRoutes(r, h)

@@ -217,7 +223,7 @@ func TestGenerateContentMixedToolSnippetAlsoTriggersFunctionCall(t *testing.T) {
 func TestStreamGenerateContentEmitsSSE(t *testing.T) {
 	h := &Handler{
 		Store:  testGeminiConfig{},
-		OpenAI: geminiOpenAISuccessStub{stream: true},
+		OpenAI: &geminiOpenAISuccessStub{stream: true},
 	}
 	r := chi.NewRouter()
 	RegisterRoutes(r, h)
@@ -251,6 +257,39 @@ func TestStreamGenerateContentEmitsSSE(t *testing.T) {
 	}
 }

+func TestGeminiProxyTranslatesInlineImageToOpenAIDataURL(t *testing.T) {
+	openAI := &geminiOpenAISuccessStub{}
+	h := &Handler{Store: testGeminiConfig{}, OpenAI: openAI}
+	r := chi.NewRouter()
+	RegisterRoutes(r, h)
+
+	body := `{"contents":[{"role":"user","parts":[{"text":"hello"},{"inlineData":{"mimeType":"image/png","data":"QUJDRA=="}}]}]}`
+	req := httptest.NewRequest(http.MethodPost, "/v1beta/models/gemini-2.5-pro:generateContent", strings.NewReader(body))
+	rec := httptest.NewRecorder()
+	r.ServeHTTP(rec, req)
+
+	if rec.Code != http.StatusOK {
+		t.Fatalf("expected 200, got %d body=%s", rec.Code, rec.Body.String())
+	}
+	messages, _ := openAI.seenReq["messages"].([]any)
+	if len(messages) != 1 {
+		t.Fatalf("expected one translated message, got %#v", openAI.seenReq)
+	}
+	msg, _ := messages[0].(map[string]any)
+	content, _ := msg["content"].([]any)
+	if len(content) != 2 {
+		t.Fatalf("expected translated content blocks, got %#v", msg)
+	}
+	imageBlock, _ := content[1].(map[string]any)
+	if strings.TrimSpace(asString(imageBlock["type"])) != "image_url" {
+		t.Fatalf("expected image_url block, got %#v", imageBlock)
+	}
+	imageURL, _ := imageBlock["image_url"].(map[string]any)
+	if !strings.HasPrefix(strings.TrimSpace(asString(imageURL["url"])), "data:image/png;base64,") {
+		t.Fatalf("expected translated data url, got %#v", imageBlock)
+	}
+}
+
 func TestGenerateContentOpenAIProxyErrorUsesGeminiEnvelope(t *testing.T) {
 	h := &Handler{
 		Store: testGeminiConfig{},
--- a/internal/adapter/openai/deps.go
+++ b/internal/adapter/openai/deps.go
@@ -18,6 +18,7 @@ type AuthResolver interface {
 type DeepSeekCaller interface {
 	CreateSession(ctx context.Context, a *auth.RequestAuth, maxAttempts int) (string, error)
 	GetPow(ctx context.Context, a *auth.RequestAuth, maxAttempts int) (string, error)
+	UploadFile(ctx context.Context, a *auth.RequestAuth, req deepseek.UploadFileRequest, maxAttempts int) (*deepseek.UploadFileResult, error)
 	CallCompletion(ctx context.Context, a *auth.RequestAuth, payload map[string]any, powResp string, maxAttempts int) (*http.Response, error)
 	DeleteSessionForToken(ctx context.Context, token string, sessionID string) (*deepseek.DeleteSessionResult, error)
 	DeleteAllSessionsForToken(ctx context.Context, token string) error
--- a/internal/adapter/openai/file_inline_upload.go
+++ b/internal/adapter/openai/file_inline_upload.go
@@ -0,0 +1,375 @@
+package openai
+
+import (
+	"context"
+	"crypto/sha256"
+	"encoding/base64"
+	"fmt"
+	"mime"
+	"net/http"
+	"net/url"
+	"path/filepath"
+	"strings"
+
+	"ds2api/internal/auth"
+	"ds2api/internal/deepseek"
+)
+
+type inlineFileUploadError struct {
+	status  int
+	message string
+	err     error
+}
+
+func (e *inlineFileUploadError) Error() string {
+	if e == nil {
+		return ""
+	}
+	if strings.TrimSpace(e.message) != "" {
+		return e.message
+	}
+	if e.err != nil {
+		return e.err.Error()
+	}
+	return "inline file processing failed"
+}
+
+type inlineUploadState struct {
+	ctx          context.Context
+	handler      *Handler
+	auth         *auth.RequestAuth
+	uploadedByID map[string]string
+}
+
+type inlineDecodedFile struct {
+	Data            []byte
+	ContentType     string
+	Filename        string
+	ReplacementType string
+}
+
+func (h *Handler) preprocessInlineFileInputs(ctx context.Context, a *auth.RequestAuth, req map[string]any) error {
+	if h == nil || h.DS == nil || len(req) == 0 {
+		return nil
+	}
+	state := &inlineUploadState{
+		ctx:          ctx,
+		handler:      h,
+		auth:         a,
+		uploadedByID: map[string]string{},
+	}
+	for _, key := range []string{"messages", "input", "attachments"} {
+		if raw, ok := req[key]; ok {
+			updated, err := state.walk(raw)
+			if err != nil {
+				return err
+			}
+			req[key] = updated
+		}
+	}
+	if refIDs := collectOpenAIRefFileIDs(req); len(refIDs) > 0 {
+		req["ref_file_ids"] = stringsToAnySlice(refIDs)
+	}
+	return nil
+}
+
+func writeOpenAIInlineFileError(w http.ResponseWriter, err error) {
+	inlineErr, ok := err.(*inlineFileUploadError)
+	if !ok || inlineErr == nil {
+		writeOpenAIError(w, http.StatusInternalServerError, "Failed to process file input.")
+		return
+	}
+	status := inlineErr.status
+	if status == 0 {
+		status = http.StatusInternalServerError
+	}
+	message := strings.TrimSpace(inlineErr.message)
+	if message == "" {
+		message = "Failed to process file input."
+	}
+	writeOpenAIError(w, status, message)
+}
+
+func (s *inlineUploadState) walk(raw any) (any, error) {
+	switch x := raw.(type) {
+	case []any:
+		out := make([]any, len(x))
+		for i, item := range x {
+			updated, err := s.walk(item)
+			if err != nil {
+				return nil, err
+			}
+			out[i] = updated
+		}
+		return out, nil
+	case map[string]any:
+		if replacement, replaced, err := s.tryUploadBlock(x); replaced || err != nil {
+			return replacement, err
+		}
+		for _, key := range []string{"messages", "input", "attachments", "content", "files", "items", "data", "source", "file", "image_url"} {
+			if nested, ok := x[key]; ok {
+				updated, err := s.walk(nested)
+				if err != nil {
+					return nil, err
+				}
+				x[key] = updated
+			}
+		}
+		return x, nil
+	default:
+		return raw, nil
+	}
+}
+
+func (s *inlineUploadState) tryUploadBlock(block map[string]any) (map[string]any, bool, error) {
+	decoded, ok, err := decodeOpenAIInlineFileBlock(block)
+	if err != nil {
+		return nil, true, &inlineFileUploadError{status: http.StatusBadRequest, message: err.Error(), err: err}
+	}
+	if !ok {
+		return nil, false, nil
+	}
+	fileID, err := s.uploadInlineFile(decoded)
+	if err != nil {
+		return nil, true, &inlineFileUploadError{status: http.StatusInternalServerError, message: "Failed to upload inline file.", err: err}
+	}
+	replacement := map[string]any{
+		"type":    decoded.ReplacementType,
+		"file_id": fileID,
+	}
+	if decoded.Filename != "" {
+		replacement["filename"] = decoded.Filename
+	}
+	if decoded.ContentType != "" {
+		replacement["mime_type"] = decoded.ContentType
+	}
+	return replacement, true, nil
+}
+
+func (s *inlineUploadState) uploadInlineFile(file inlineDecodedFile) (string, error) {
+	sum := sha256.Sum256(append(append([]byte(file.ContentType+"\x00"+file.Filename+"\x00"), file.Data...)))
+	cacheKey := fmt.Sprintf("%x", sum[:])
+	if fileID, ok := s.uploadedByID[cacheKey]; ok && strings.TrimSpace(fileID) != "" {
+		return fileID, nil
+	}
+	contentType := strings.TrimSpace(file.ContentType)
+	if contentType == "" {
+		contentType = http.DetectContentType(file.Data)
+	}
+	result, err := s.handler.DS.UploadFile(s.ctx, s.auth, deepseek.UploadFileRequest{
+		Filename:    file.Filename,
+		ContentType: contentType,
+		Data:        file.Data,
+	}, 3)
+	if err != nil {
+		return "", err
+	}
+	fileID := strings.TrimSpace(result.ID)
+	if fileID == "" {
+		return "", fmt.Errorf("upload succeeded without file id")
+	}
+	s.uploadedByID[cacheKey] = fileID
+	return fileID, nil
+}
+
+func decodeOpenAIInlineFileBlock(block map[string]any) (inlineDecodedFile, bool, error) {
+	if block == nil {
+		return inlineDecodedFile{}, false, nil
+	}
+	if strings.TrimSpace(asString(block["file_id"])) != "" {
+		return inlineDecodedFile{}, false, nil
+	}
+	if nested, ok := block["file"].(map[string]any); ok {
+		decoded, matched, err := decodeOpenAIInlineFileBlock(nested)
+		if err != nil || !matched {
+			return decoded, matched, err
+		}
+		if decoded.Filename == "" {
+			decoded.Filename = pickInlineFilename(block, decoded.ContentType, defaultInlinePrefix(decoded.ReplacementType))
+		}
+		return decoded, true, nil
+	}
+	blockType := strings.ToLower(strings.TrimSpace(asString(block["type"])))
+	if raw, matched := extractInlineImageDataURL(block); matched {
+		data, contentType, err := decodeInlinePayload(raw, contentTypeFromMap(block))
+		if err != nil {
+			return inlineDecodedFile{}, true, fmt.Errorf("invalid image input")
+		}
+		return inlineDecodedFile{
+			Data:            data,
+			ContentType:     contentType,
+			Filename:        pickInlineFilename(block, contentType, "image"),
+			ReplacementType: "input_image",
+		}, true, nil
+	}
+	if raw, matched := extractInlineFilePayload(block, blockType); matched {
+		data, contentType, err := decodeInlinePayload(raw, contentTypeFromMap(block))
+		if err != nil {
+			return inlineDecodedFile{}, true, fmt.Errorf("invalid file input")
+		}
+		return inlineDecodedFile{
+			Data:            data,
+			ContentType:     contentType,
+			Filename:        pickInlineFilename(block, contentType, defaultInlinePrefix(blockType)),
+			ReplacementType: "input_file",
+		}, true, nil
+	}
+	return inlineDecodedFile{}, false, nil
+}
+
+func extractInlineImageDataURL(block map[string]any) (string, bool) {
+	imageURL := block["image_url"]
+	switch x := imageURL.(type) {
+	case string:
+		if isDataURL(x) {
+			return strings.TrimSpace(x), true
+		}
+	case map[string]any:
+		if raw := strings.TrimSpace(asString(x["url"])); isDataURL(raw) {
+			return raw, true
+		}
+	}
+	if raw := strings.TrimSpace(asString(block["url"])); isDataURL(raw) {
+		return raw, true
+	}
+	return "", false
+}
+
+func extractInlineFilePayload(block map[string]any, blockType string) (string, bool) {
+	for _, value := range []any{block["file_data"], block["base64"], block["data"]} {
+		if raw := strings.TrimSpace(asString(value)); raw != "" {
+			if strings.Contains(blockType, "file") || block["file_data"] != nil || block["filename"] != nil || block["file_name"] != nil || block["name"] != nil {
+				return raw, true
+			}
+		}
+	}
+	return "", false
+}
+
+func decodeInlinePayload(raw string, explicitContentType string) ([]byte, string, error) {
+	raw = strings.TrimSpace(raw)
+	if raw == "" {
+		return nil, "", fmt.Errorf("empty payload")
+	}
+	if isDataURL(raw) {
+		return decodeDataURL(raw, explicitContentType)
+	}
+	decoded, err := decodeBase64Flexible(raw)
+	if err != nil {
+		return nil, "", err
+	}
+	contentType := strings.TrimSpace(explicitContentType)
+	if contentType == "" && len(decoded) > 0 {
+		contentType = http.DetectContentType(decoded)
+	}
+	return decoded, contentType, nil
+}
+
+func decodeDataURL(raw string, explicitContentType string) ([]byte, string, error) {
+	raw = strings.TrimSpace(raw)
+	if !isDataURL(raw) {
+		return nil, "", fmt.Errorf("unsupported data url")
+	}
+	header, payload, ok := strings.Cut(raw, ",")
+	if !ok {
+		return nil, "", fmt.Errorf("invalid data url")
+	}
+	meta := strings.TrimSpace(strings.TrimPrefix(header, "data:"))
+	contentType := strings.TrimSpace(explicitContentType)
+	if contentType == "" {
+		contentType = "application/octet-stream"
+		if meta != "" {
+			parts := strings.Split(meta, ";")
+			if len(parts) > 0 && strings.TrimSpace(parts[0]) != "" {
+				contentType = strings.TrimSpace(parts[0])
+			}
+		}
+	}
+	if strings.Contains(strings.ToLower(meta), ";base64") {
+		decoded, err := decodeBase64Flexible(payload)
+		if err != nil {
+			return nil, "", err
+		}
+		return decoded, contentType, nil
+	}
+	decoded, err := url.QueryUnescape(payload)
+	if err != nil {
+		return nil, "", err
+	}
+	return []byte(decoded), contentType, nil
+}
+
+func decodeBase64Flexible(raw string) ([]byte, error) {
+	raw = strings.TrimSpace(raw)
+	for _, enc := range []*base64.Encoding{base64.StdEncoding, base64.RawStdEncoding, base64.URLEncoding, base64.RawURLEncoding} {
+		decoded, err := enc.DecodeString(raw)
+		if err == nil {
+			return decoded, nil
+		}
+	}
+	return nil, fmt.Errorf("invalid base64 payload")
+}
+
+func contentTypeFromMap(block map[string]any) string {
+	for _, value := range []any{block["mime_type"], block["mimeType"], block["content_type"], block["contentType"], block["media_type"], block["mediaType"]} {
+		if contentType := strings.TrimSpace(asString(value)); contentType != "" {
+			return contentType
+		}
+	}
+	if imageURL, ok := block["image_url"].(map[string]any); ok {
+		for _, value := range []any{imageURL["mime_type"], imageURL["mimeType"], imageURL["content_type"], imageURL["contentType"]} {
+			if contentType := strings.TrimSpace(asString(value)); contentType != "" {
+				return contentType
+			}
+		}
+	}
+	return ""
+}
+
+func pickInlineFilename(block map[string]any, contentType string, prefix string) string {
+	for _, value := range []any{block["filename"], block["file_name"], block["name"]} {
+		if name := strings.TrimSpace(asString(value)); name != "" {
+			return filepath.Base(name)
+		}
+	}
+	if prefix == "" {
+		prefix = "upload"
+	}
+	ext := ".bin"
+	if parsedType := strings.TrimSpace(contentType); parsedType != "" {
+		if comma := strings.Index(parsedType, ";"); comma >= 0 {
+			parsedType = strings.TrimSpace(parsedType[:comma])
+		}
+		if exts, err := mime.ExtensionsByType(parsedType); err == nil && len(exts) > 0 && strings.TrimSpace(exts[0]) != "" {
+			ext = exts[0]
+		}
+	}
+	return prefix + ext
+}
+
+func defaultInlinePrefix(blockType string) string {
+	blockType = strings.ToLower(strings.TrimSpace(blockType))
+	if strings.Contains(blockType, "image") {
+		return "image"
+	}
+	return "upload"
+}
+
+func isDataURL(raw string) bool {
+	return strings.HasPrefix(strings.ToLower(strings.TrimSpace(raw)), "data:")
+}
+
+func stringsToAnySlice(items []string) []any {
+	out := make([]any, 0, len(items))
+	for _, item := range items {
+		trimmed := strings.TrimSpace(item)
+		if trimmed == "" {
+			continue
+		}
+		out = append(out, trimmed)
+	}
+	if len(out) == 0 {
+		return nil
+	}
+	return out
+}
--- a/internal/adapter/openai/file_inline_upload_test.go
+++ b/internal/adapter/openai/file_inline_upload_test.go
@@ -0,0 +1,274 @@
+package openai
+
+import (
+	"context"
+	"encoding/json"
+	"errors"
+	"net/http"
+	"net/http/httptest"
+	"strings"
+	"testing"
+
+	"github.com/go-chi/chi/v5"
+
+	"ds2api/internal/auth"
+	"ds2api/internal/deepseek"
+)
+
+type inlineUploadDSStub struct {
+	uploadCalls    []deepseek.UploadFileRequest
+	lastCtx        context.Context
+	completionReq  map[string]any
+	createSession  string
+	uploadErr      error
+	completionResp *http.Response
+}
+
+func (m *inlineUploadDSStub) CreateSession(_ context.Context, _ *auth.RequestAuth, _ int) (string, error) {
+	if strings.TrimSpace(m.createSession) == "" {
+		return "session-id", nil
+	}
+	return m.createSession, nil
+}
+
+func (m *inlineUploadDSStub) GetPow(_ context.Context, _ *auth.RequestAuth, _ int) (string, error) {
+	return "pow", nil
+}
+
+func (m *inlineUploadDSStub) UploadFile(ctx context.Context, _ *auth.RequestAuth, req deepseek.UploadFileRequest, _ int) (*deepseek.UploadFileResult, error) {
+	m.lastCtx = ctx
+	m.uploadCalls = append(m.uploadCalls, req)
+	if m.uploadErr != nil {
+		return nil, m.uploadErr
+	}
+	return &deepseek.UploadFileResult{
+		ID:       "file-inline-1",
+		Filename: req.Filename,
+		Bytes:    int64(len(req.Data)),
+		Status:   "uploaded",
+		Purpose:  req.Purpose,
+	}, nil
+}
+
+func (m *inlineUploadDSStub) CallCompletion(_ context.Context, _ *auth.RequestAuth, payload map[string]any, _ string, _ int) (*http.Response, error) {
+	m.completionReq = payload
+	if m.completionResp != nil {
+		return m.completionResp, nil
+	}
+	return makeOpenAISSEHTTPResponse(
+		`data: {"p":"response/content","v":"ok"}`,
+		`data: [DONE]`,
+	), nil
+}
+
+func (m *inlineUploadDSStub) DeleteSessionForToken(_ context.Context, _ string, _ string) (*deepseek.DeleteSessionResult, error) {
+	return &deepseek.DeleteSessionResult{Success: true}, nil
+}
+
+func (m *inlineUploadDSStub) DeleteAllSessionsForToken(_ context.Context, _ string) error {
+	return nil
+}
+
+func TestPreprocessInlineFileInputsReplacesDataURLAndCollectsRefFileIDs(t *testing.T) {
+	ds := &inlineUploadDSStub{}
+	h := &Handler{DS: ds}
+	req := map[string]any{
+		"messages": []any{
+			map[string]any{
+				"role": "user",
+				"content": []any{
+					map[string]any{
+						"type":      "image_url",
+						"image_url": map[string]any{"url": "data:image/png;base64,QUJDRA=="},
+					},
+				},
+			},
+		},
+	}
+	ctx, cancel := context.WithCancel(context.Background())
+	defer cancel()
+
+	if err := h.preprocessInlineFileInputs(ctx, &auth.RequestAuth{DeepSeekToken: "token"}, req); err != nil {
+		t.Fatalf("preprocess failed: %v", err)
+	}
+	if len(ds.uploadCalls) != 1 {
+		t.Fatalf("expected 1 upload, got %d", len(ds.uploadCalls))
+	}
+	if ds.lastCtx != ctx {
+		t.Fatalf("expected upload to use request context")
+	}
+	if ds.uploadCalls[0].ContentType != "image/png" {
+		t.Fatalf("expected image/png, got %q", ds.uploadCalls[0].ContentType)
+	}
+	if ds.uploadCalls[0].Filename != "image.png" {
+		t.Fatalf("expected inferred filename image.png, got %q", ds.uploadCalls[0].Filename)
+	}
+	messages, _ := req["messages"].([]any)
+	first, _ := messages[0].(map[string]any)
+	content, _ := first["content"].([]any)
+	block, _ := content[0].(map[string]any)
+	if block["type"] != "input_image" {
+		t.Fatalf("expected input_image replacement, got %#v", block)
+	}
+	if block["file_id"] != "file-inline-1" {
+		t.Fatalf("expected file-inline-1 replacement id, got %#v", block)
+	}
+	refIDs, _ := req["ref_file_ids"].([]any)
+	if len(refIDs) != 1 || refIDs[0] != "file-inline-1" {
+		t.Fatalf("unexpected ref_file_ids: %#v", req["ref_file_ids"])
+	}
+}
+
+func TestPreprocessInlineFileInputsDeduplicatesIdenticalPayloads(t *testing.T) {
+	ds := &inlineUploadDSStub{}
+	h := &Handler{DS: ds}
+	req := map[string]any{
+		"messages": []any{
+			map[string]any{
+				"role": "user",
+				"content": []any{
+					map[string]any{"type": "image_url", "image_url": map[string]any{"url": "data:image/png;base64,QUJDRA=="}},
+					map[string]any{"type": "image_url", "image_url": map[string]any{"url": "data:image/png;base64,QUJDRA=="}},
+				},
+			},
+		},
+	}
+
+	if err := h.preprocessInlineFileInputs(context.Background(), &auth.RequestAuth{DeepSeekToken: "token"}, req); err != nil {
+		t.Fatalf("preprocess failed: %v", err)
+	}
+	if len(ds.uploadCalls) != 1 {
+		t.Fatalf("expected deduplicated single upload, got %d", len(ds.uploadCalls))
+	}
+	refIDs, _ := req["ref_file_ids"].([]any)
+	if len(refIDs) != 1 || refIDs[0] != "file-inline-1" {
+		t.Fatalf("unexpected ref_file_ids after dedupe: %#v", req["ref_file_ids"])
+	}
+}
+
+func TestChatCompletionsUploadsInlineFilesBeforeCompletion(t *testing.T) {
+	ds := &inlineUploadDSStub{}
+	h := &Handler{Store: mockOpenAIConfig{wideInput: true}, Auth: streamStatusAuthStub{}, DS: ds}
+	reqBody := `{"model":"deepseek-chat","messages":[{"role":"user","content":[{"type":"input_text","text":"hi"},{"type":"image_url","image_url":{"url":"data:image/png;base64,QUJDRA=="}}]}],"stream":false}`
+	req := httptest.NewRequest(http.MethodPost, "/v1/chat/completions", strings.NewReader(reqBody))
+	req.Header.Set("Authorization", "Bearer direct-token")
+	req.Header.Set("Content-Type", "application/json")
+	rec := httptest.NewRecorder()
+
+	h.ChatCompletions(rec, req)
+
+	if rec.Code != http.StatusOK {
+		t.Fatalf("expected 200, got %d body=%s", rec.Code, rec.Body.String())
+	}
+	if len(ds.uploadCalls) != 1 {
+		t.Fatalf("expected 1 upload call, got %d", len(ds.uploadCalls))
+	}
+	if ds.completionReq == nil {
+		t.Fatal("expected completion payload to be captured")
+	}
+	refIDs, _ := ds.completionReq["ref_file_ids"].([]any)
+	if len(refIDs) != 1 || refIDs[0] != "file-inline-1" {
+		t.Fatalf("unexpected completion ref_file_ids: %#v", ds.completionReq["ref_file_ids"])
+	}
+}
+
+func TestResponsesUploadsInlineFilesBeforeCompletion(t *testing.T) {
+	ds := &inlineUploadDSStub{}
+	h := &Handler{Store: mockOpenAIConfig{wideInput: true}, Auth: streamStatusAuthStub{}, DS: ds}
+	r := chi.NewRouter()
+	RegisterRoutes(r, h)
+	reqBody := `{"model":"deepseek-chat","input":[{"role":"user","content":[{"type":"input_text","text":"hi"},{"type":"input_image","image_url":{"url":"data:image/png;base64,QUJDRA=="}}]}],"stream":false}`
+	req := httptest.NewRequest(http.MethodPost, "/v1/responses", strings.NewReader(reqBody))
+	req.Header.Set("Authorization", "Bearer direct-token")
+	req.Header.Set("Content-Type", "application/json")
+	rec := httptest.NewRecorder()
+
+	r.ServeHTTP(rec, req)
+
+	if rec.Code != http.StatusOK {
+		t.Fatalf("expected 200, got %d body=%s", rec.Code, rec.Body.String())
+	}
+	if len(ds.uploadCalls) != 1 {
+		t.Fatalf("expected 1 upload call, got %d", len(ds.uploadCalls))
+	}
+	refIDs, _ := ds.completionReq["ref_file_ids"].([]any)
+	if len(refIDs) != 1 || refIDs[0] != "file-inline-1" {
+		t.Fatalf("unexpected completion ref_file_ids: %#v", ds.completionReq["ref_file_ids"])
+	}
+}
+
+func TestChatCompletionsInlineUploadFailureReturnsBadRequest(t *testing.T) {
+	ds := &inlineUploadDSStub{}
+	h := &Handler{Store: mockOpenAIConfig{wideInput: true}, Auth: streamStatusAuthStub{}, DS: ds}
+	reqBody := `{"model":"deepseek-chat","messages":[{"role":"user","content":[{"type":"image_url","image_url":{"url":"data:image/png;base64,%%%"}}]}],"stream":false}`
+	req := httptest.NewRequest(http.MethodPost, "/v1/chat/completions", strings.NewReader(reqBody))
+	req.Header.Set("Authorization", "Bearer direct-token")
+	req.Header.Set("Content-Type", "application/json")
+	rec := httptest.NewRecorder()
+
+	h.ChatCompletions(rec, req)
+
+	if rec.Code != http.StatusBadRequest {
+		t.Fatalf("expected 400, got %d body=%s", rec.Code, rec.Body.String())
+	}
+	if ds.completionReq != nil {
+		t.Fatalf("did not expect completion call on upload decode error")
+	}
+}
+
+func TestResponsesInlineUploadFailureReturnsInternalServerError(t *testing.T) {
+	ds := &inlineUploadDSStub{uploadErr: errors.New("boom")}
+	h := &Handler{Store: mockOpenAIConfig{wideInput: true}, Auth: streamStatusAuthStub{}, DS: ds}
+	r := chi.NewRouter()
+	RegisterRoutes(r, h)
+	reqBody := `{"model":"deepseek-chat","input":[{"role":"user","content":[{"type":"image_url","image_url":{"url":"data:image/png;base64,QUJDRA=="}}]}],"stream":false}`
+	req := httptest.NewRequest(http.MethodPost, "/v1/responses", strings.NewReader(reqBody))
+	req.Header.Set("Authorization", "Bearer direct-token")
+	req.Header.Set("Content-Type", "application/json")
+	rec := httptest.NewRecorder()
+
+	r.ServeHTTP(rec, req)
+
+	if rec.Code != http.StatusInternalServerError {
+		t.Fatalf("expected 500, got %d body=%s", rec.Code, rec.Body.String())
+	}
+	if ds.completionReq != nil {
+		t.Fatalf("did not expect completion call after upload failure")
+	}
+}
+
+func TestVercelPrepareUploadsInlineFilesBeforeLeasePayload(t *testing.T) {
+	t.Setenv("VERCEL", "1")
+	t.Setenv("DS2API_VERCEL_INTERNAL_SECRET", "stream-secret")
+	ds := &inlineUploadDSStub{}
+	h := &Handler{Store: mockOpenAIConfig{wideInput: true}, Auth: streamStatusAuthStub{}, DS: ds}
+	r := chi.NewRouter()
+	RegisterRoutes(r, h)
+	reqBody := `{"model":"deepseek-chat","messages":[{"role":"user","content":[{"type":"input_text","text":"hi"},{"type":"image_url","image_url":{"url":"data:image/png;base64,QUJDRA=="}}]}],"stream":true}`
+	req := httptest.NewRequest(http.MethodPost, "/v1/chat/completions?__stream_prepare=1", strings.NewReader(reqBody))
+	req.Header.Set("Authorization", "Bearer direct-token")
+	req.Header.Set("X-Ds2-Internal-Token", "stream-secret")
+	req.Header.Set("Content-Type", "application/json")
+	rec := httptest.NewRecorder()
+
+	r.ServeHTTP(rec, req)
+
+	if rec.Code != http.StatusOK {
+		t.Fatalf("expected 200, got %d body=%s", rec.Code, rec.Body.String())
+	}
+	if len(ds.uploadCalls) != 1 {
+		t.Fatalf("expected 1 upload call, got %d", len(ds.uploadCalls))
+	}
+	var out map[string]any
+	if err := json.Unmarshal(rec.Body.Bytes(), &out); err != nil {
+		t.Fatalf("decode response failed: %v body=%s", err, rec.Body.String())
+	}
+	payload, _ := out["payload"].(map[string]any)
+	if payload == nil {
+		t.Fatalf("expected payload in prepare response, got %#v", out)
+	}
+	refIDs, _ := payload["ref_file_ids"].([]any)
+	if len(refIDs) != 1 || refIDs[0] != "file-inline-1" {
+		t.Fatalf("unexpected payload ref_file_ids: %#v", payload["ref_file_ids"])
+	}
+}
--- a/internal/adapter/openai/file_refs.go
+++ b/internal/adapter/openai/file_refs.go
@@ -0,0 +1,73 @@
+package openai
+
+import "strings"
+
+func collectOpenAIRefFileIDs(req map[string]any) []string {
+	if len(req) == 0 {
+		return nil
+	}
+	out := make([]string, 0, 4)
+	seen := map[string]struct{}{}
+	for _, raw := range []any{
+		req["ref_file_ids"],
+		req["file_ids"],
+		req["attachments"],
+		req["messages"],
+		req["input"],
+	} {
+		appendOpenAIRefFileIDs(&out, seen, raw)
+	}
+	if len(out) == 0 {
+		return nil
+	}
+	return out
+}
+
+func appendOpenAIRefFileIDs(out *[]string, seen map[string]struct{}, raw any) {
+	switch x := raw.(type) {
+	case string:
+		addOpenAIRefFileID(out, seen, x)
+	case []string:
+		for _, item := range x {
+			addOpenAIRefFileID(out, seen, item)
+		}
+	case []any:
+		for _, item := range x {
+			appendOpenAIRefFileIDs(out, seen, item)
+		}
+	case map[string]any:
+		if fileID := strings.TrimSpace(asString(x["file_id"])); fileID != "" {
+			addOpenAIRefFileID(out, seen, fileID)
+		}
+		if strings.Contains(strings.ToLower(strings.TrimSpace(asString(x["type"]))), "file") {
+			if fileID := strings.TrimSpace(asString(x["id"])); fileID != "" {
+				addOpenAIRefFileID(out, seen, fileID)
+			}
+		}
+		if fileMap, ok := x["file"].(map[string]any); ok {
+			if fileID := strings.TrimSpace(asString(fileMap["file_id"])); fileID != "" {
+				addOpenAIRefFileID(out, seen, fileID)
+			}
+			if fileID := strings.TrimSpace(asString(fileMap["id"])); fileID != "" {
+				addOpenAIRefFileID(out, seen, fileID)
+			}
+		}
+		for _, key := range []string{"ref_file_ids", "file_ids", "attachments", "messages", "input", "content", "files", "items", "data", "source"} {
+			if nested, ok := x[key]; ok {
+				appendOpenAIRefFileIDs(out, seen, nested)
+			}
+		}
+	}
+}
+
+func addOpenAIRefFileID(out *[]string, seen map[string]struct{}, fileID string) {
+	fileID = strings.TrimSpace(fileID)
+	if fileID == "" {
+		return
+	}
+	if _, ok := seen[fileID]; ok {
+		return
+	}
+	seen[fileID] = struct{}{}
+	*out = append(*out, fileID)
+}
--- a/internal/adapter/openai/files_route_test.go
+++ b/internal/adapter/openai/files_route_test.go
@@ -0,0 +1,156 @@
+package openai
+
+import (
+	"bytes"
+	"context"
+	"encoding/json"
+	"errors"
+	"mime/multipart"
+	"net/http"
+	"net/http/httptest"
+	"testing"
+
+	"github.com/go-chi/chi/v5"
+
+	"ds2api/internal/auth"
+	"ds2api/internal/deepseek"
+)
+
+type filesRouteDSStub struct {
+	lastReq deepseek.UploadFileRequest
+	upload  *deepseek.UploadFileResult
+	err     error
+}
+
+func (m *filesRouteDSStub) CreateSession(_ context.Context, _ *auth.RequestAuth, _ int) (string, error) {
+	return "", nil
+}
+
+func (m *filesRouteDSStub) GetPow(_ context.Context, _ *auth.RequestAuth, _ int) (string, error) {
+	return "", nil
+}
+
+func (m *filesRouteDSStub) UploadFile(_ context.Context, _ *auth.RequestAuth, req deepseek.UploadFileRequest, _ int) (*deepseek.UploadFileResult, error) {
+	m.lastReq = req
+	if m.err != nil {
+		return nil, m.err
+	}
+	if m.upload != nil {
+		return m.upload, nil
+	}
+	return &deepseek.UploadFileResult{ID: "file-123", Filename: req.Filename, Bytes: int64(len(req.Data)), Purpose: req.Purpose, Status: "uploaded"}, nil
+}
+
+func (m *filesRouteDSStub) CallCompletion(_ context.Context, _ *auth.RequestAuth, _ map[string]any, _ string, _ int) (*http.Response, error) {
+	return nil, errors.New("not implemented")
+}
+
+func (m *filesRouteDSStub) DeleteSessionForToken(_ context.Context, _ string, _ string) (*deepseek.DeleteSessionResult, error) {
+	return &deepseek.DeleteSessionResult{Success: true}, nil
+}
+
+func (m *filesRouteDSStub) DeleteAllSessionsForToken(_ context.Context, _ string) error {
+	return nil
+}
+
+func newMultipartUploadRequest(t *testing.T, purpose string, filename string, data []byte) *http.Request {
+	t.Helper()
+	var body bytes.Buffer
+	writer := multipart.NewWriter(&body)
+	if purpose != "" {
+		if err := writer.WriteField("purpose", purpose); err != nil {
+			t.Fatalf("write purpose failed: %v", err)
+		}
+	}
+	part, err := writer.CreateFormFile("file", filename)
+	if err != nil {
+		t.Fatalf("create form file failed: %v", err)
+	}
+	if _, err := part.Write(data); err != nil {
+		t.Fatalf("write file failed: %v", err)
+	}
+	if err := writer.Close(); err != nil {
+		t.Fatalf("close writer failed: %v", err)
+	}
+	req := httptest.NewRequest(http.MethodPost, "/v1/files", &body)
+	req.Header.Set("Authorization", "Bearer direct-token")
+	req.Header.Set("Content-Type", writer.FormDataContentType())
+	return req
+}
+
+func TestFilesRouteUploadSuccess(t *testing.T) {
+	ds := &filesRouteDSStub{}
+	h := &Handler{Store: mockOpenAIConfig{wideInput: true}, Auth: streamStatusAuthStub{}, DS: ds}
+	r := chi.NewRouter()
+	RegisterRoutes(r, h)
+
+	req := newMultipartUploadRequest(t, "assistants", "notes.txt", []byte("hello world"))
+	rec := httptest.NewRecorder()
+	r.ServeHTTP(rec, req)
+
+	if rec.Code != http.StatusOK {
+		t.Fatalf("expected 200, got %d body=%s", rec.Code, rec.Body.String())
+	}
+	if ds.lastReq.Filename != "notes.txt" {
+		t.Fatalf("expected filename notes.txt, got %q", ds.lastReq.Filename)
+	}
+	if ds.lastReq.Purpose != "assistants" {
+		t.Fatalf("expected purpose assistants, got %q", ds.lastReq.Purpose)
+	}
+	if string(ds.lastReq.Data) != "hello world" {
+		t.Fatalf("unexpected uploaded data: %q", string(ds.lastReq.Data))
+	}
+	var out map[string]any
+	if err := json.Unmarshal(rec.Body.Bytes(), &out); err != nil {
+		t.Fatalf("decode response failed: %v body=%s", err, rec.Body.String())
+	}
+	if out["object"] != "file" {
+		t.Fatalf("expected file object, got %#v", out)
+	}
+	if out["id"] != "file-123" {
+		t.Fatalf("expected file id file-123, got %#v", out["id"])
+	}
+	if out["filename"] != "notes.txt" {
+		t.Fatalf("expected filename notes.txt, got %#v", out["filename"])
+	}
+}
+
+func TestFilesRouteRejectsNonMultipart(t *testing.T) {
+	h := &Handler{Store: mockOpenAIConfig{wideInput: true}, Auth: streamStatusAuthStub{}, DS: &filesRouteDSStub{}}
+	r := chi.NewRouter()
+	RegisterRoutes(r, h)
+
+	req := httptest.NewRequest(http.MethodPost, "/v1/files", bytes.NewBufferString(`{"purpose":"assistants"}`))
+	req.Header.Set("Authorization", "Bearer direct-token")
+	req.Header.Set("Content-Type", "application/json")
+	rec := httptest.NewRecorder()
+	r.ServeHTTP(rec, req)
+
+	if rec.Code != http.StatusBadRequest {
+		t.Fatalf("expected 400, got %d body=%s", rec.Code, rec.Body.String())
+	}
+}
+
+func TestFilesRouteRequiresFileField(t *testing.T) {
+	h := &Handler{Store: mockOpenAIConfig{wideInput: true}, Auth: streamStatusAuthStub{}, DS: &filesRouteDSStub{}}
+	r := chi.NewRouter()
+	RegisterRoutes(r, h)
+
+	var body bytes.Buffer
+	writer := multipart.NewWriter(&body)
+	if err := writer.WriteField("purpose", "assistants"); err != nil {
+		t.Fatalf("write field failed: %v", err)
+	}
+	if err := writer.Close(); err != nil {
+		t.Fatalf("close writer failed: %v", err)
+	}
+	req := httptest.NewRequest(http.MethodPost, "/v1/files", &body)
+	req.Header.Set("Authorization", "Bearer direct-token")
+	req.Header.Set("Content-Type", writer.FormDataContentType())
+	rec := httptest.NewRecorder()
+	r.ServeHTTP(rec, req)
+
+	if rec.Code != http.StatusBadRequest {
+		t.Fatalf("expected 400, got %d body=%s", rec.Code, rec.Body.String())
+	}
+}
--- a/internal/adapter/openai/handler_chat.go
+++ b/internal/adapter/openai/handler_chat.go
@@ -48,6 +48,10 @@ func (h *Handler) ChatCompletions(w http.ResponseWriter, r *http.Request) {
 		writeOpenAIError(w, http.StatusBadRequest, "invalid json")
 		return
 	}
+	if err := h.preprocessInlineFileInputs(r.Context(), a, req); err != nil {
+		writeOpenAIInlineFileError(w, err)
+		return
+	}
 	stdReq, err := normalizeOpenAIChatRequest(h.Store, req, requestTraceID(r))
 	if err != nil {
 		writeOpenAIError(w, http.StatusBadRequest, err.Error())
--- a/internal/adapter/openai/handler_chat_auto_delete_test.go
+++ b/internal/adapter/openai/handler_chat_auto_delete_test.go
@@ -27,6 +27,10 @@ func (m *autoDeleteModeDSStub) GetPow(_ context.Context, _ *auth.RequestAuth, _
 	return "pow", nil
 }

+func (m *autoDeleteModeDSStub) UploadFile(_ context.Context, _ *auth.RequestAuth, _ deepseek.UploadFileRequest, _ int) (*deepseek.UploadFileResult, error) {
+	return &deepseek.UploadFileResult{ID: "file-id", Filename: "file.txt", Bytes: 1, Status: "uploaded"}, nil
+}
+
 func (m *autoDeleteModeDSStub) CallCompletion(_ context.Context, _ *auth.RequestAuth, _ map[string]any, _ string, _ int) (*http.Response, error) {
 	return m.resp, nil
 }
--- a/internal/adapter/openai/handler_files.go
+++ b/internal/adapter/openai/handler_files.go
@@ -0,0 +1,90 @@
+package openai
+
+import (
+	"io"
+	"net/http"
+	"strings"
+	"time"
+
+	"ds2api/internal/auth"
+	"ds2api/internal/deepseek"
+)
+
+const openAIUploadMaxMemory = 32 << 20
+
+func (h *Handler) UploadFile(w http.ResponseWriter, r *http.Request) {
+	a, err := h.Auth.Determine(r)
+	if err != nil {
+		status := http.StatusUnauthorized
+		detail := err.Error()
+		if err == auth.ErrNoAccount {
+			status = http.StatusTooManyRequests
+		}
+		writeOpenAIError(w, status, detail)
+		return
+	}
+	defer h.Auth.Release(a)
+	if !strings.HasPrefix(strings.ToLower(strings.TrimSpace(r.Header.Get("Content-Type"))), "multipart/form-data") {
+		writeOpenAIError(w, http.StatusBadRequest, "content-type must be multipart/form-data")
+		return
+	}
+	if err := r.ParseMultipartForm(openAIUploadMaxMemory); err != nil {
+		writeOpenAIError(w, http.StatusBadRequest, "invalid multipart form")
+		return
+	}
+	if r.MultipartForm != nil {
+		defer r.MultipartForm.RemoveAll()
+	}
+	r = r.WithContext(auth.WithAuth(r.Context(), a))
+	file, header, err := r.FormFile("file")
+	if err != nil {
+		writeOpenAIError(w, http.StatusBadRequest, "file is required")
+		return
+	}
+	defer file.Close()
+	data, err := io.ReadAll(file)
+	if err != nil {
+		writeOpenAIError(w, http.StatusBadRequest, "failed to read uploaded file")
+		return
+	}
+	contentType := strings.TrimSpace(header.Header.Get("Content-Type"))
+	if contentType == "" && len(data) > 0 {
+		contentType = http.DetectContentType(data)
+	}
+	result, err := h.DS.UploadFile(r.Context(), a, deepseek.UploadFileRequest{
+		Filename:    header.Filename,
+		ContentType: contentType,
+		Purpose:     strings.TrimSpace(r.FormValue("purpose")),
+		Data:        data,
+	}, 3)
+	if err != nil {
+		writeOpenAIError(w, http.StatusInternalServerError, "Failed to upload file.")
+		return
+	}
+	writeJSON(w, http.StatusOK, buildOpenAIFileObject(result))
+}
+
+func buildOpenAIFileObject(result *deepseek.UploadFileResult) map[string]any {
+	if result == nil {
+		return map[string]any{
+			"id":             "",
+			"object":         "file",
+			"bytes":          0,
+			"created_at":     time.Now().Unix(),
+			"filename":       "",
+			"purpose":        "",
+			"status":         "uploaded",
+			"status_details": nil,
+		}
+	}
+	return map[string]any{
+		"id":             result.ID,
+		"object":         "file",
+		"bytes":          result.Bytes,
+		"created_at":     time.Now().Unix(),
+		"filename":       result.Filename,
+		"purpose":        result.Purpose,
+		"status":         result.Status,
+		"status_details": nil,
+	}
+}
--- a/internal/adapter/openai/handler_routes.go
+++ b/internal/adapter/openai/handler_routes.go
@@ -46,6 +46,7 @@ func RegisterRoutes(r chi.Router, h *Handler) {
 	r.Post("/v1/chat/completions", h.ChatCompletions)
 	r.Post("/v1/responses", h.Responses)
 	r.Get("/v1/responses/{response_id}", h.GetResponseByID)
+	r.Post("/v1/files", h.UploadFile)
 	r.Post("/v1/embeddings", h.Embeddings)
 }

--- a/internal/adapter/openai/responses_embeddings_test.go
+++ b/internal/adapter/openai/responses_embeddings_test.go
@@ -156,6 +156,33 @@ func TestNormalizeResponsesInputAsMessagesFunctionCallItemPreservesConcatenatedA
 	}
 }

+func TestCollectOpenAIRefFileIDs(t *testing.T) {
+	got := collectOpenAIRefFileIDs(map[string]any{
+		"ref_file_ids": []any{"file-top", "file-dup"},
+		"attachments": []any{
+			map[string]any{"file_id": "file-attachment"},
+		},
+		"input": []any{
+			map[string]any{
+				"type": "message",
+				"content": []any{
+					map[string]any{"type": "input_file", "file_id": "file-input"},
+					map[string]any{"type": "input_file", "id": "file-dup"},
+				},
+			},
+		},
+	})
+	want := []string{"file-top", "file-dup", "file-attachment", "file-input"}
+	if len(got) != len(want) {
+		t.Fatalf("expected %d file ids, got %#v", len(want), got)
+	}
+	for i, id := range want {
+		if got[i] != id {
+			t.Fatalf("unexpected file ids at %d: got=%#v want=%#v", i, got, want)
+		}
+	}
+}
+
 func TestExtractEmbeddingInputs(t *testing.T) {
 	got := extractEmbeddingInputs([]any{"a", "b"})
 	if len(got) != 2 || got[0] != "a" || got[1] != "b" {
--- a/internal/adapter/openai/responses_handler.go
+++ b/internal/adapter/openai/responses_handler.go
@@ -70,6 +70,10 @@ func (h *Handler) Responses(w http.ResponseWriter, r *http.Request) {
 		writeOpenAIError(w, http.StatusBadRequest, "invalid json")
 		return
 	}
+	if err := h.preprocessInlineFileInputs(r.Context(), a, req); err != nil {
+		writeOpenAIInlineFileError(w, err)
+		return
+	}
 	traceID := requestTraceID(r)
 	stdReq, err := normalizeOpenAIResponsesRequest(h.Store, req, traceID)
 	if err != nil {
--- a/internal/adapter/openai/standard_request.go
+++ b/internal/adapter/openai/standard_request.go
@@ -27,6 +27,7 @@ func normalizeOpenAIChatRequest(store ConfigReader, req map[string]any, traceID
 	finalPrompt, toolNames := buildOpenAIFinalPromptWithPolicy(messagesRaw, req["tools"], traceID, toolPolicy, thinkingEnabled)
 	toolNames = ensureToolDetectionEnabled(toolNames, req["tools"])
 	passThrough := collectOpenAIChatPassThrough(req)
+	refFileIDs := collectOpenAIRefFileIDs(req)

 	return util.StandardRequest{
 		Surface:        "openai_chat",
@@ -40,6 +41,7 @@ func normalizeOpenAIChatRequest(store ConfigReader, req map[string]any, traceID
 		Stream:         util.ToBool(req["stream"]),
 		Thinking:       thinkingEnabled,
 		Search:         searchEnabled,
+		RefFileIDs:     refFileIDs,
 		PassThrough:    passThrough,
 	}, nil
 }
@@ -80,6 +82,7 @@ func normalizeOpenAIResponsesRequest(store ConfigReader, req map[string]any, tra
 		toolPolicy.Allowed = namesToSet(toolNames)
 	}
 	passThrough := collectOpenAIChatPassThrough(req)
+	refFileIDs := collectOpenAIRefFileIDs(req)

 	return util.StandardRequest{
 		Surface:        "openai_responses",
@@ -93,6 +96,7 @@ func normalizeOpenAIResponsesRequest(store ConfigReader, req map[string]any, tra
 		Stream:         util.ToBool(req["stream"]),
 		Thinking:       thinkingEnabled,
 		Search:         searchEnabled,
+		RefFileIDs:     refFileIDs,
 		PassThrough:    passThrough,
 	}, nil
 }
--- a/internal/adapter/openai/standard_request_test.go
+++ b/internal/adapter/openai/standard_request_test.go
@@ -41,6 +41,36 @@ func TestNormalizeOpenAIChatRequest(t *testing.T) {
 	}
 }

+func TestNormalizeOpenAIChatRequestCollectsRefFileIDs(t *testing.T) {
+	store := newEmptyStoreForNormalizeTest(t)
+	req := map[string]any{
+		"model": "gpt-5-codex",
+		"messages": []any{
+			map[string]any{
+				"role": "user",
+				"content": []any{
+					map[string]any{"type": "input_text", "text": "hello"},
+					map[string]any{"type": "input_file", "file_id": "file-msg"},
+				},
+			},
+		},
+		"attachments": []any{
+			map[string]any{"file_id": "file-attachment"},
+		},
+		"ref_file_ids": []any{"file-top", "file-attachment"},
+	}
+	n, err := normalizeOpenAIChatRequest(store, req, "")
+	if err != nil {
+		t.Fatalf("normalize failed: %v", err)
+	}
+	if len(n.RefFileIDs) != 3 {
+		t.Fatalf("expected 3 distinct file ids, got %#v", n.RefFileIDs)
+	}
+	if n.RefFileIDs[0] != "file-top" || n.RefFileIDs[1] != "file-attachment" || n.RefFileIDs[2] != "file-msg" {
+		t.Fatalf("unexpected file ids: %#v", n.RefFileIDs)
+	}
+}
+
 func TestNormalizeOpenAIResponsesRequestInput(t *testing.T) {
 	store := newEmptyStoreForNormalizeTest(t)
 	req := map[string]any{
--- a/internal/adapter/openai/stream_status_test.go
+++ b/internal/adapter/openai/stream_status_test.go
@@ -50,6 +50,10 @@ func (m streamStatusDSStub) GetPow(_ context.Context, _ *auth.RequestAuth, _ int
 	return "pow", nil
 }

+func (m streamStatusDSStub) UploadFile(_ context.Context, _ *auth.RequestAuth, _ deepseek.UploadFileRequest, _ int) (*deepseek.UploadFileResult, error) {
+	return &deepseek.UploadFileResult{ID: "file-id", Filename: "file.txt", Bytes: 1, Status: "uploaded"}, nil
+}
+
 func (m streamStatusDSStub) CallCompletion(_ context.Context, _ *auth.RequestAuth, _ map[string]any, _ string, _ int) (*http.Response, error) {
 	return m.resp, nil
 }
--- a/internal/adapter/openai/vercel_stream.go
+++ b/internal/adapter/openai/vercel_stream.go
@@ -52,6 +52,10 @@ func (h *Handler) handleVercelStreamPrepare(w http.ResponseWriter, r *http.Reque
 		writeOpenAIError(w, http.StatusBadRequest, "invalid json")
 		return
 	}
+	if err := h.preprocessInlineFileInputs(r.Context(), a, req); err != nil {
+		writeOpenAIInlineFileError(w, err)
+		return
+	}
 	if !util.ToBool(req["stream"]) {
 		writeOpenAIError(w, http.StatusBadRequest, "stream must be true")
 		return