cpu-diffusers/backend.proto

syntax = "proto3";

option go_package = "github.com/go-skynet/LocalAI/pkg/grpc/proto";
option java_multiple_files = true;
option java_package = "io.skynet.localai.backend";
option java_outer_classname = "LocalAIBackend";

package backend;

service Backend {
  rpc Health(HealthMessage) returns (Reply) {}
  rpc Free(HealthMessage) returns (Result) {}
  rpc Predict(PredictOptions) returns (Reply) {}
  rpc LoadModel(ModelOptions) returns (Result) {}
  rpc PredictStream(PredictOptions) returns (stream Reply) {}
  rpc Embedding(PredictOptions) returns (EmbeddingResult) {}
  rpc GenerateImage(GenerateImageRequest) returns (Result) {}
  rpc GenerateVideo(GenerateVideoRequest) returns (Result) {}
  rpc AudioTranscription(TranscriptRequest) returns (TranscriptResult) {}
  rpc AudioTranscriptionStream(TranscriptRequest) returns (stream TranscriptStreamResponse) {}
  rpc TTS(TTSRequest) returns (Result) {}
  rpc TTSStream(TTSRequest) returns (stream Reply) {}
  rpc SoundGeneration(SoundGenerationRequest) returns (Result) {}
  rpc TokenizeString(PredictOptions) returns (TokenizationResponse) {}
  rpc Status(HealthMessage) returns (StatusResponse) {}
  rpc Detect(DetectOptions) returns (DetectResponse) {}
  rpc FaceVerify(FaceVerifyRequest) returns (FaceVerifyResponse) {}
  rpc FaceAnalyze(FaceAnalyzeRequest) returns (FaceAnalyzeResponse) {}
  rpc VoiceVerify(VoiceVerifyRequest) returns (VoiceVerifyResponse) {}
  rpc VoiceAnalyze(VoiceAnalyzeRequest) returns (VoiceAnalyzeResponse) {}
  rpc VoiceEmbed(VoiceEmbedRequest) returns (VoiceEmbedResponse) {}

  rpc StoresSet(StoresSetOptions) returns (Result) {}
  rpc StoresDelete(StoresDeleteOptions) returns (Result) {}
  rpc StoresGet(StoresGetOptions) returns (StoresGetResult) {}
  rpc StoresFind(StoresFindOptions) returns (StoresFindResult) {}

  rpc Rerank(RerankRequest) returns (RerankResult) {}

  rpc GetMetrics(MetricsRequest) returns (MetricsResponse);

  rpc VAD(VADRequest) returns (VADResponse) {}

  rpc Diarize(DiarizeRequest) returns (DiarizeResponse) {}

  rpc AudioEncode(AudioEncodeRequest) returns (AudioEncodeResult) {}
  rpc AudioDecode(AudioDecodeRequest) returns (AudioDecodeResult) {}

  rpc AudioTransform(AudioTransformRequest) returns (AudioTransformResult) {}
  rpc AudioTransformStream(stream AudioTransformFrameRequest) returns (stream AudioTransformFrameResponse) {}
  // AudioToAudioStream is the bidirectional any-to-any S2S RPC. Backends
  // that load a speech-to-speech model consume input audio frames and emit
  // interleaved audio + transcript + tool-call deltas as typed events.
  // Backends without S2S support return UNIMPLEMENTED.
  rpc AudioToAudioStream(stream AudioToAudioRequest) returns (stream AudioToAudioResponse) {}

  rpc ModelMetadata(ModelOptions) returns (ModelMetadataResponse) {}

  // Fine-tuning RPCs
  rpc StartFineTune(FineTuneRequest) returns (FineTuneJobResult) {}
  rpc FineTuneProgress(FineTuneProgressRequest) returns (stream FineTuneProgressUpdate) {}
  rpc StopFineTune(FineTuneStopRequest) returns (Result) {}
  rpc ListCheckpoints(ListCheckpointsRequest) returns (ListCheckpointsResponse) {}
  rpc ExportModel(ExportModelRequest) returns (Result) {}

  // Quantization RPCs
  rpc StartQuantization(QuantizationRequest) returns (QuantizationJobResult) {}
  rpc QuantizationProgress(QuantizationProgressRequest) returns (stream QuantizationProgressUpdate) {}
  rpc StopQuantization(QuantizationStopRequest) returns (Result) {}

}

// Define the empty request
message MetricsRequest {}

message MetricsResponse {
  int32 slot_id = 1;
  string prompt_json_for_slot = 2;  // Stores the prompt as a JSON string.
  float tokens_per_second = 3;
  int32 tokens_generated = 4;
  int32 prompt_tokens_processed = 5;
}

message RerankRequest {
  string query = 1;
  repeated string documents = 2;
  int32 top_n = 3;
}

message RerankResult {
  Usage usage = 1;
  repeated DocumentResult results = 2;
}

message Usage {
  int32 total_tokens = 1;
  int32 prompt_tokens = 2;
}

message DocumentResult {
  int32 index = 1;
  string text = 2;
  float relevance_score = 3;
}

message StoresKey {
  repeated float Floats = 1;
}

message StoresValue {
  bytes Bytes = 1;
}

message StoresSetOptions {
  repeated StoresKey Keys = 1;
  repeated StoresValue Values = 2;
}

message StoresDeleteOptions {
  repeated StoresKey Keys = 1;
}

message StoresGetOptions {
  repeated StoresKey Keys = 1;
}

message StoresGetResult {
  repeated StoresKey Keys = 1;
  repeated StoresValue Values = 2;
}

message StoresFindOptions {
  StoresKey Key = 1;
  int32 TopK = 2;
}

message StoresFindResult {
  repeated StoresKey Keys = 1;
  repeated StoresValue Values = 2;
  repeated float Similarities = 3;
}

message HealthMessage {}

// The request message containing the user's name.
message PredictOptions {
  string Prompt = 1;
  int32 Seed = 2;
  int32 Threads = 3;
  int32 Tokens = 4;
  int32 TopK = 5;
  int32 Repeat = 6;
  int32 Batch = 7;
  int32 NKeep = 8;
  float Temperature = 9;
  float Penalty = 10;
  bool F16KV = 11;
  bool DebugMode = 12;
  repeated string StopPrompts = 13;
  bool IgnoreEOS = 14;
  float TailFreeSamplingZ = 15;
  float TypicalP = 16;
  float FrequencyPenalty = 17;
  float PresencePenalty = 18;
  int32 Mirostat = 19;
  float MirostatETA = 20;
  float MirostatTAU = 21;
  bool PenalizeNL = 22;
  string LogitBias = 23;
  bool MLock = 25;
  bool MMap = 26;
  bool PromptCacheAll = 27;
  bool PromptCacheRO = 28;
  string Grammar = 29;
  string MainGPU = 30;
  string TensorSplit = 31;
  float TopP = 32;
  string PromptCachePath = 33;
  bool Debug = 34;
  repeated int32 EmbeddingTokens = 35;
  string Embeddings = 36;
  float RopeFreqBase = 37;
  float RopeFreqScale = 38;
  float NegativePromptScale = 39;
  string NegativePrompt = 40;
  int32 NDraft = 41;
  repeated string Images = 42;
  bool UseTokenizerTemplate = 43;
  repeated Message Messages = 44;
  repeated string Videos = 45;
  repeated string Audios = 46;
  string CorrelationId = 47;
  string Tools = 48;  // JSON array of available tools/functions for tool calling
  string ToolChoice = 49;  // JSON string or object specifying tool choice behavior
  int32 Logprobs = 50;  // Number of top logprobs to return (maps to OpenAI logprobs parameter)
  int32 TopLogprobs = 51;  // Number of top logprobs to return per token (maps to OpenAI top_logprobs parameter)
  map<string, string> Metadata = 52;  // Generic per-request metadata (e.g., enable_thinking)
  float MinP = 53;  // Minimum probability sampling threshold (0.0 = disabled)
}

// ToolCallDelta represents an incremental tool call update from the C++ parser.
// Used for both streaming (partial diffs) and non-streaming (final tool calls).
message ToolCallDelta {
  int32 index = 1;           // tool call index (0-based)
  string id = 2;             // tool call ID (e.g., "call_abc123")
  string name = 3;           // function name (set on first appearance)
  string arguments = 4;      // arguments chunk (incremental in streaming, full in non-streaming)
}

// ChatDelta represents incremental content/reasoning/tool_call updates parsed by the C++ backend.
message ChatDelta {
  string content = 1;                       // content text delta
  string reasoning_content = 2;             // reasoning/thinking text delta
  repeated ToolCallDelta tool_calls = 3;    // tool call deltas
}

// The response message containing the result
message Reply {
  bytes message = 1;
  int32 tokens = 2;
  int32 prompt_tokens = 3;
  double timing_prompt_processing = 4;
  double timing_token_generation = 5;
  bytes audio = 6;
  bytes logprobs = 7;  // JSON-encoded logprobs data matching OpenAI format
  repeated ChatDelta chat_deltas = 8;       // Parsed chat deltas from C++ autoparser (streaming + non-streaming)
}

message GrammarTrigger {
  string word = 1;
}

message ModelOptions {
  string Model = 1;
  int32 ContextSize = 2;
  int32 Seed = 3;
  int32 NBatch = 4;
  bool F16Memory = 5;
  bool MLock = 6;
  bool MMap = 7;
  bool VocabOnly = 8;
  bool LowVRAM = 9;
  bool Embeddings = 10;
  bool NUMA = 11;
  int32 NGPULayers = 12;
  string MainGPU = 13;
  string TensorSplit = 14;
  int32 Threads = 15;
  float RopeFreqBase = 17;
  float RopeFreqScale = 18;
  float RMSNormEps = 19;
  int32 NGQA = 20;
  string ModelFile = 21;



  // Diffusers
  string PipelineType = 26;
  string SchedulerType = 27;
  bool CUDA = 28;
  float CFGScale = 29;
  bool IMG2IMG = 30;
  string CLIPModel = 31;
  string CLIPSubfolder = 32;
  int32 CLIPSkip = 33;
  string ControlNet = 48;

  string Tokenizer = 34;

  // LLM (llama.cpp)
  string LoraBase = 35;
  string LoraAdapter = 36;
  float LoraScale = 42;

  bool NoMulMatQ = 37;
  string DraftModel = 39;

  string AudioPath = 38;

  // vllm
  string Quantization = 40;
  float  GPUMemoryUtilization = 50;
  bool   TrustRemoteCode = 51;
  bool   EnforceEager = 52;
  int32  SwapSpace = 53;
  int32  MaxModelLen = 54;
  int32  TensorParallelSize = 55;
  string LoadFormat = 58;
  bool   DisableLogStatus = 66;
  string DType = 67;
  int32  LimitImagePerPrompt = 68;
  int32  LimitVideoPerPrompt = 69;
  int32  LimitAudioPerPrompt = 70;

  string MMProj = 41;

  string RopeScaling = 43;
  float YarnExtFactor = 44;
  float YarnAttnFactor = 45;
  float YarnBetaFast = 46;
  float YarnBetaSlow = 47;

  string Type = 49;

  string FlashAttention = 56;
  bool NoKVOffload = 57;

  string ModelPath = 59;

  repeated string LoraAdapters = 60;
  repeated float LoraScales = 61;

  repeated string Options = 62;

  string CacheTypeKey = 63;
  string CacheTypeValue = 64;

  repeated GrammarTrigger GrammarTriggers = 65;

  bool Reranking = 71;

  repeated string Overrides = 72;

  // EngineArgs carries a JSON-encoded map of backend-native engine arguments
  // applied verbatim to the backend's engine constructor (e.g. vLLM AsyncEngineArgs).
  // Unknown keys produce an error at LoadModel time.
  string EngineArgs = 73;
}

message Result {
  string message = 1;
  bool success = 2;
}

message EmbeddingResult {
  repeated float embeddings = 1;
}

message TranscriptRequest {
  string dst = 2;
  string language = 3;
  uint32 threads = 4;
  bool translate = 5;
  bool diarize = 6;
  string prompt = 7;
  float temperature = 8;
  repeated string timestamp_granularities = 9;
  bool stream = 10;
}

message TranscriptResult {
  repeated TranscriptSegment segments = 1;
  string text = 2;
  string language = 3;
  float duration = 4;
}

message TranscriptStreamResponse {
  string delta = 1;
  TranscriptResult final_result = 2;
}

message TranscriptWord {
  int64 start = 1;
  int64 end = 2;
  string text = 3;
}

message TranscriptSegment {
  int32 id = 1;
  int64 start = 2;
  int64 end = 3;
  string text = 4;
  repeated int32 tokens = 5;
  string speaker = 6;
  repeated TranscriptWord words = 7;
}

message GenerateImageRequest {
  int32 height = 1;
  int32 width = 2;
  int32 step = 4;
  int32 seed = 5;
  string positive_prompt = 6;
  string negative_prompt = 7;
  string dst = 8;
  string src = 9;

  // Diffusers
  string EnableParameters = 10;
  int32 CLIPSkip = 11;

  // Reference images for models that support them (e.g., Flux Kontext)
  repeated string ref_images = 12;
}

message GenerateVideoRequest {
  string prompt = 1;
  string negative_prompt = 2;  // Negative prompt for video generation
  string start_image = 3;  // Path or base64 encoded image for the start frame
  string end_image = 4;    // Path or base64 encoded image for the end frame
  int32 width = 5;
  int32 height = 6;
  int32 num_frames = 7;    // Number of frames to generate
  int32 fps = 8;          // Frames per second
  int32 seed = 9;
  float cfg_scale = 10;    // Classifier-free guidance scale
  int32 step = 11;         // Number of inference steps
  string dst = 12;        // Output path for the generated video
}

message TTSRequest {
  string text = 1;
  string model = 2;
  string dst = 3;
  string voice = 4;
  optional string language = 5;
}

message VADRequest {
  repeated float audio = 1;
}

message VADSegment {
  float start = 1;
  float end = 2;
}

message VADResponse {
  repeated VADSegment segments = 1;
}

// --- Speaker diarization messages ---
//
// Pure speaker diarization: "who spoke when". Returns time-stamped segments
// labelled with cluster IDs (the same string for the same speaker across
// segments). Some backends (e.g. vibevoice.cpp) produce diarization as a
// by-product of ASR and may also fill in `text` per segment; backends with a
// dedicated diarization pipeline (e.g. sherpa-onnx pyannote) leave `text`
// empty and emit only the segmentation.

message DiarizeRequest {
  string dst = 1;                      // path to audio file (HTTP layer materialises uploads to a temp file)
  uint32 threads = 2;
  string language = 3;                 // optional; only meaningful for transcription-bundling backends
  int32  num_speakers = 4;             // exact speaker count if known (>0 forces); 0 = auto
  int32  min_speakers = 5;             // hint when auto-detecting; 0 = unset
  int32  max_speakers = 6;             // hint when auto-detecting; 0 = unset
  float  clustering_threshold = 7;     // distance threshold when num_speakers unknown; 0 = backend default
  float  min_duration_on = 8;          // discard segments shorter than this (seconds); 0 = backend default
  float  min_duration_off = 9;         // merge gaps shorter than this (seconds); 0 = backend default
  bool   include_text = 10;            // when the backend can emit per-segment transcript for free, ask it to populate `text`
}

message DiarizeSegment {
  int32  id = 1;
  float  start = 2;                    // seconds
  float  end = 3;                      // seconds
  string speaker = 4;                  // backend-emitted speaker label (e.g. "0", "SPEAKER_00")
  string text = 5;                     // optional per-segment transcript (empty unless include_text and supported)
}

message DiarizeResponse {
  repeated DiarizeSegment segments = 1;
  int32  num_speakers = 2;             // count of distinct speaker labels in `segments`
  float  duration = 3;                 // total audio duration in seconds (0 if unknown)
  string language = 4;                 // optional, when the backend bundles transcription
}

message SoundGenerationRequest {
  string text = 1;
  string model = 2;
  string dst = 3;
  optional float duration = 4;
  optional float temperature = 5;
  optional bool sample = 6;
  optional string src = 7;
  optional int32 src_divisor = 8;
  optional bool think = 9;
  optional string caption = 10;
  optional string lyrics = 11;
  optional int32 bpm = 12;
  optional string keyscale = 13;
  optional string language = 14;
  optional string timesignature = 15;
  optional bool instrumental = 17;
}

message TokenizationResponse {
  int32 length = 1;
  repeated int32 tokens = 2;
}

message MemoryUsageData {
  uint64 total = 1;
  map<string, uint64> breakdown = 2;
}

message StatusResponse {
  enum State {
    UNINITIALIZED = 0;
    BUSY = 1;
    READY = 2;
    ERROR = -1;
  }
  State state = 1;
  MemoryUsageData memory = 2;
}

message Message {
  string role = 1;
  string content = 2;
  // Optional fields for OpenAI-compatible message format
  string name = 3;                    // Tool name (for tool messages)
  string tool_call_id = 4;            // Tool call ID (for tool messages)
  string reasoning_content = 5;       // Reasoning content (for thinking models)
  string tool_calls = 6;              // Tool calls as JSON string (for assistant messages with tool calls)
}

message DetectOptions {
  string src = 1;
  string prompt = 2;           // Text prompt (for SAM 3 PCS mode)
  repeated float points = 3;   // Point coordinates as [x1, y1, label1, x2, y2, label2, ...] (label: 1=pos, 0=neg)
  repeated float boxes = 4;    // Box coordinates as [x1, y1, x2, y2, ...]
  float threshold = 5;         // Detection confidence threshold
}

message Detection {
  float x = 1;
  float y = 2;
  float width = 3;
  float height = 4;
  float confidence = 5;
  string class_name = 6;
  bytes mask = 7;              // PNG-encoded binary segmentation mask
}

message DetectResponse {
  repeated Detection Detections = 1;
}

// --- Face recognition messages ---

message FacialArea {
  float x = 1;
  float y = 2;
  float w = 3;
  float h = 4;
}

message FaceVerifyRequest {
  string img1 = 1;              // base64-encoded image
  string img2 = 2;              // base64-encoded image
  float  threshold = 3;         // cosine-distance threshold; 0 = use backend default
  bool   anti_spoofing = 4;     // run MiniFASNet liveness on each image; failed liveness forces verified=false
}

message FaceVerifyResponse {
  bool       verified = 1;
  float      distance = 2;      // 1 - cosine_similarity
  float      threshold = 3;
  float      confidence = 4;    // 0-100
  string     model = 5;         // e.g. "buffalo_l"
  FacialArea img1_area = 6;
  FacialArea img2_area = 7;
  float      processing_time_ms = 8;
  bool       img1_is_real = 9;          // anti-spoofing result when enabled
  float      img1_antispoof_score = 10;
  bool       img2_is_real = 11;
  float      img2_antispoof_score = 12;
}

message FaceAnalyzeRequest {
  string          img = 1;          // base64-encoded image
  repeated string actions = 2;      // subset of ["age","gender","emotion","race"]; empty = all-supported
  bool            anti_spoofing = 3;
}

message FaceAnalysis {
  FacialArea         region = 1;
  float              face_confidence = 2;
  float              age = 3;
  string             dominant_gender = 4;   // "Man" | "Woman"
  map<string, float> gender = 5;
  string             dominant_emotion = 6;  // reserved; empty in MVP
  map<string, float> emotion = 7;
  string             dominant_race = 8;     // not populated
  map<string, float> race = 9;
  bool               is_real = 10;          // anti-spoofing result when enabled
  float              antispoof_score = 11;
}

message FaceAnalyzeResponse {
  repeated FaceAnalysis faces = 1;
}

// --- Voice (speaker) recognition messages ---
//
// Analogous to the Face* messages above, but for speaker biometrics.
// Audio fields accept a filesystem path (same convention as
// TranscriptRequest.dst). The HTTP layer materialises base64 / URL /
// data-URI inputs to a temp file before calling the gRPC backend.

message VoiceVerifyRequest {
  string audio1 = 1;            // path to first audio clip
  string audio2 = 2;            // path to second audio clip
  float  threshold = 3;         // cosine-distance threshold; 0 = use backend default
  bool   anti_spoofing = 4;     // reserved for future AASIST bolt-on
}

message VoiceVerifyResponse {
  bool   verified = 1;
  float  distance = 2;          // 1 - cosine_similarity
  float  threshold = 3;
  float  confidence = 4;        // 0-100
  string model = 5;             // e.g. "speechbrain/spkrec-ecapa-voxceleb"
  float  processing_time_ms = 6;
}

message VoiceAnalyzeRequest {
  string          audio = 1;        // path to audio clip
  repeated string actions = 2;      // subset of ["age","gender","emotion"]; empty = all-supported
}

message VoiceAnalysis {
  float              start = 1;          // segment start time in seconds (0 if single-utterance)
  float              end = 2;            // segment end time in seconds
  float              age = 3;
  string             dominant_gender = 4;
  map<string, float> gender = 5;
  string             dominant_emotion = 6;
  map<string, float> emotion = 7;
}

message VoiceAnalyzeResponse {
  repeated VoiceAnalysis segments = 1;
}

message VoiceEmbedRequest {
  string audio = 1;              // path to audio clip
}

message VoiceEmbedResponse {
  repeated float embedding = 1;
  string         model = 2;
}

message ToolFormatMarkers {
  string format_type = 1;           // "json_native", "tag_with_json", "tag_with_tagged"

  // Tool section markers
  string section_start = 2;         // e.g., "<tool_call>", "[TOOL_CALLS]"
  string section_end = 3;           // e.g., "</tool_call>"
  string per_call_start = 4;        // e.g., "<|tool_call_begin|>"
  string per_call_end = 5;          // e.g., "<|tool_call_end|>"

  // Function name markers (TAG_WITH_JSON / TAG_WITH_TAGGED)
  string func_name_prefix = 6;     // e.g., "<function="
  string func_name_suffix = 7;     // e.g., ">"
  string func_close = 8;           // e.g., "</function>"

  // Argument markers (TAG_WITH_TAGGED)
  string arg_name_prefix = 9;      // e.g., "<param="
  string arg_name_suffix = 10;     // e.g., ">"
  string arg_value_prefix = 11;
  string arg_value_suffix = 12;    // e.g., "</param>"
  string arg_separator = 13;       // e.g., "\n"

  // JSON format fields (JSON_NATIVE)
  string name_field = 14;          // e.g., "name"
  string args_field = 15;          // e.g., "arguments"
  string id_field = 16;            // e.g., "id"
  bool fun_name_is_key = 17;
  bool tools_array_wrapped = 18;
  reserved 19;

  // Reasoning markers
  string reasoning_start = 20;     // e.g., "<think>"
  string reasoning_end = 21;       // e.g., "</think>"

  // Content markers
  string content_start = 22;
  string content_end = 23;

  // Args wrapper markers
  string args_start = 24;          // e.g., "<args>"
  string args_end = 25;            // e.g., "</args>"

  // JSON parameter ordering
  string function_field = 26;      // e.g., "function" (wrapper key in JSON)
  repeated string parameter_order = 27;

  // Generated ID field (alternative field name for generated IDs)
  string gen_id_field = 28;        // e.g., "call_id"

  // Call ID markers (position and delimiters for tool call IDs)
  string call_id_position = 29;    // "none", "pre_func_name", "between_func_and_args", "post_args"
  string call_id_prefix = 30;      // e.g., "[CALL_ID]"
  string call_id_suffix = 31;      // e.g., ""
}

message AudioEncodeRequest {
  bytes pcm_data = 1;
  int32 sample_rate = 2;
  int32 channels = 3;
  map<string, string> options = 4;
}

message AudioEncodeResult {
  repeated bytes frames = 1;
  int32 sample_rate = 2;
  int32 samples_per_frame = 3;
}

message AudioDecodeRequest {
  repeated bytes frames = 1;
  map<string, string> options = 2;
}

message AudioDecodeResult {
  bytes pcm_data = 1;
  int32 sample_rate = 2;
  int32 samples_per_frame = 3;
}

// Generic audio transform: an audio-in, audio-out operation, optionally
// conditioned on a second reference signal. Concrete transforms include
// AEC + noise suppression + dereverberation (LocalVQE), voice conversion
// (reference = target speaker), pitch shifting, etc.
message AudioTransformRequest {
  string audio_path = 1;             // required, primary input file path
  string reference_path = 2;         // optional auxiliary; empty => zero-fill
  string dst = 3;                    // required, output file path
  map<string, string> params = 4;    // backend-specific tuning
}

message AudioTransformResult {
  string dst = 1;
  int32  sample_rate = 2;
  int32  samples = 3;
  bool   reference_provided = 4;
}

// Bidirectional streaming audio transform. The first message MUST carry a
// Config; subsequent messages carry Frames. A second Config mid-stream
// resets streaming state before the next frame.
message AudioTransformFrameRequest {
  oneof payload {
    AudioTransformStreamConfig config = 1;
    AudioTransformFrame        frame  = 2;
  }
}

message AudioTransformStreamConfig {
  enum SampleFormat {
    F32_LE = 0;
    S16_LE = 1;
  }
  SampleFormat sample_format = 1;
  int32 sample_rate = 2;             // 0 => backend default
  int32 frame_samples = 3;           // 0 => backend default
  map<string, string> params = 4;
  bool reset = 5;                    // reset streaming state before next frame
}

message AudioTransformFrame {
  bytes audio_pcm = 1;               // frame_samples samples in stream's format
  bytes reference_pcm = 2;           // empty => zero-fill (silent reference)
}

message AudioTransformFrameResponse {
  bytes pcm = 1;
  int64 frame_index = 2;
}

// === AudioToAudioStream messages =========================================
//
// Bidirectional stream between the LocalAI core and an any-to-any audio
// model. The client opens the stream with a Config payload, then alternates
// Frame (input audio) and Control (turn boundaries, function-call results,
// session updates) payloads. The server streams back typed events: audio
// frames carry PCM in `pcm`; transcript / tool-call deltas carry JSON in
// `meta`; the stream ends with a `response.done` (success) or `error` event.

message AudioToAudioRequest {
  oneof payload {
    AudioToAudioConfig  config  = 1;
    AudioToAudioFrame   frame   = 2;
    AudioToAudioControl control = 3;
  }
}

message AudioToAudioConfig {
  // PCM format for client→server audio. 0 => backend default
  // (16 kHz for the LFM2-Audio Conformer encoder).
  int32 input_sample_rate = 1;
  // Preferred server→client audio rate. 0 => backend default
  // (24 kHz for the LFM2-Audio vocoder).
  int32 output_sample_rate = 2;
  // Optional system prompt override. Empty => backend chooses based on
  // mode (e.g. "Respond with interleaved text and audio.").
  string system_prompt = 3;
  // Optional baked-voice id. Models that only ship a fixed set of
  // voices (e.g. LFM2-Audio: us_male/us_female/uk_male/uk_female) match
  // this against their voice table; an empty string keeps the default.
  string voice = 4;
  // JSON-encoded array of tool definitions in OpenAI Chat Completions
  // format. Empty => no tools.
  string tools = 5;
  // Free-form sampling / decoding parameters (temperature, top_k,
  // max_new_tokens, audio_top_k, etc).
  map<string, string> params = 6;
  // True => reset any session-scoped state before processing further
  // frames on this stream. The first Config implicitly resets.
  bool reset = 7;
}

message AudioToAudioFrame {
  // Raw PCM s16le mono at config.input_sample_rate. Empty pcm + end_of_input
  // is a valid "user finished speaking" marker without trailing audio.
  bytes pcm = 1;
  // Marks the last frame of a user turn. The backend may begin emitting
  // a response immediately after seeing this.
  bool end_of_input = 2;
}

message AudioToAudioControl {
  // Free-form control event names. Initial set:
  //   "input_audio_buffer.commit"     — user finished speaking
  //   "response.cancel"               — abort in-flight generation
  //   "conversation.item.create"      — inject a non-audio item (e.g.
  //                                     function_call_output as JSON in
  //                                     `payload`)
  //   "session.update"                — re-configure mid-stream
  string event = 1;
  // Event-specific JSON payload.
  bytes payload = 2;
}

message AudioToAudioResponse {
  // Event identifies what this frame carries. Mirrors the OpenAI Realtime
  // API server-event names where applicable. Initial set:
  //   "response.audio.delta"
  //   "response.audio_transcript.delta"
  //   "response.function_call_arguments.delta"
  //   "response.function_call_arguments.done"
  //   "response.done"
  //   "error"
  string event = 1;
  // Populated when event = response.audio.delta.
  bytes pcm = 2;
  // Populated alongside pcm to identify its rate. 0 => same as the
  // session's negotiated output_sample_rate.
  int32 sample_rate = 3;
  // JSON payload for non-PCM events (transcript chunk, tool args, error
  // body).
  bytes meta = 4;
  // Monotonic per-stream counter, useful for client reordering and
  // debugging.
  int64 sequence = 5;
}

message ModelMetadataResponse {
  bool supports_thinking = 1;
  string rendered_template = 2;  // The rendered chat template with enable_thinking=true (empty if not applicable)
  ToolFormatMarkers tool_format = 3;  // Auto-detected tool format markers from differential template analysis
  string media_marker = 4;  // Marker the backend expects in the prompt for each multimodal input (images/audio/video). Empty when the backend does not use a marker.
}

// Fine-tuning messages

message FineTuneRequest {
  // Model identification
  string model = 1;                       // HF model name or local path
  string training_type = 2;              // "lora", "loha", "lokr", "full" — what parameters to train
  string training_method = 3;            // "sft", "dpo", "grpo", "rloo", "reward", "kto", "orpo", "network_training"

  // Adapter config (universal across LoRA/LoHa/LoKr for LLM + diffusion)
  int32 adapter_rank = 10;               // LoRA rank (r), default 16
  int32 adapter_alpha = 11;              // scaling factor, default 16
  float adapter_dropout = 12;            // default 0.0
  repeated string target_modules = 13;   // layer names to adapt

  // Universal training hyperparameters
  float learning_rate = 20;              // default 2e-4
  int32 num_epochs = 21;                 // default 3
  int32 batch_size = 22;                 // default 2
  int32 gradient_accumulation_steps = 23; // default 4
  int32 warmup_steps = 24;              // default 5
  int32 max_steps = 25;                 // 0 = use epochs
  int32 save_steps = 26;               // 0 = only save final
  float weight_decay = 27;             // default 0.01
  bool gradient_checkpointing = 28;
  string optimizer = 29;               // adamw_8bit, adamw, sgd, adafactor, prodigy
  int32 seed = 30;                     // default 3407
  string mixed_precision = 31;        // fp16, bf16, fp8, no

  // Dataset
  string dataset_source = 40;          // HF dataset ID, local file/dir path
  string dataset_split = 41;           // train, test, etc.

  // Output
  string output_dir = 50;
  string job_id = 51;                  // client-assigned or auto-generated

  // Resume training from a checkpoint
  string resume_from_checkpoint = 55;  // path to checkpoint dir to resume from

  // Backend-specific AND method-specific extensibility
  map<string, string> extra_options = 60;
}

message FineTuneJobResult {
  string job_id = 1;
  bool success = 2;
  string message = 3;
}

message FineTuneProgressRequest {
  string job_id = 1;
}

message FineTuneProgressUpdate {
  string job_id = 1;
  int32 current_step = 2;
  int32 total_steps = 3;
  float current_epoch = 4;
  float total_epochs = 5;
  float loss = 6;
  float learning_rate = 7;
  float grad_norm = 8;
  float eval_loss = 9;
  float eta_seconds = 10;
  float progress_percent = 11;
  string status = 12;                  // queued, caching, loading_model, loading_dataset, training, saving, completed, failed, stopped
  string message = 13;
  string checkpoint_path = 14;        // set when a checkpoint is saved
  string sample_path = 15;           // set when a sample is generated (video/image backends)
  map<string, float> extra_metrics = 16; // method-specific metrics
}

message FineTuneStopRequest {
  string job_id = 1;
  bool save_checkpoint = 2;
}

message ListCheckpointsRequest {
  string output_dir = 1;
}

message ListCheckpointsResponse {
  repeated CheckpointInfo checkpoints = 1;
}

message CheckpointInfo {
  string path = 1;
  int32 step = 2;
  float epoch = 3;
  float loss = 4;
  string created_at = 5;
}

message ExportModelRequest {
  string checkpoint_path = 1;
  string output_path = 2;
  string export_format = 3;           // lora, loha, lokr, merged_16bit, merged_4bit, gguf, diffusers
  string quantization_method = 4;     // for GGUF: q4_k_m, q5_k_m, q8_0, f16, etc.
  string model = 5;                   // base model name (for merge operations)
  map<string, string> extra_options = 6;
}

// Quantization messages

message QuantizationRequest {
  string model = 1;                      // HF model name or local path
  string quantization_type = 2;          // q4_k_m, q5_k_m, q8_0, f16, etc.
  string output_dir = 3;                 // where to write output files
  string job_id = 4;                     // client-assigned job ID
  map<string, string> extra_options = 5; // hf_token, custom flags, etc.
}

message QuantizationJobResult {
  string job_id = 1;
  bool success = 2;
  string message = 3;
}

message QuantizationProgressRequest {
  string job_id = 1;
}

message QuantizationProgressUpdate {
  string job_id = 1;
  float progress_percent = 2;
  string status = 3;                     // queued, downloading, converting, quantizing, completed, failed, stopped
  string message = 4;
  string output_file = 5;               // set when completed — path to the output GGUF file
  map<string, float> extra_metrics = 6;  // e.g. file_size_mb, compression_ratio
}

message QuantizationStopRequest {
  string job_id = 1;
}