{ "openapi": "3.0.3", "info": { "title": "infinity_v3", "description": "infinity_v3 inference server", "contact": { "name": "infinity authors" }, "license": { "name": "Infinity Restricted MIT License 1.0 (see scope and retained licenses)" }, "version": "3.0.0" }, "paths": { "/decode": { "post": { "tags": [ "infinity_v3" ], "summary": "Decode input ids", "operationId": "decode", "requestBody": { "content": { "application/json": { "schema": { "$ref": "#/components/schemas/DecodeRequest" } } }, "required": true }, "responses": { "200": { "description": "Decoded ids", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/DecodeResponse" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/DecodeResponse" } } } }, "400": { "description": "Batch is empty", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Batch is empty", "error_type": "empty" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } }, "413": { "description": "Batch size error", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Batch size error", "error_type": "validation" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } }, "422": { "description": "Tokenization error", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Tokenization error", "error_type": "tokenizer" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } } } } }, "/embed": { "post": { "tags": [ "infinity_v3" ], "summary": "Get Embeddings. Returns a 424 status code if the model is not an embedding model.", "operationId": "embed", "requestBody": { "content": { "application/json": { "schema": { "$ref": "#/components/schemas/EmbedRequest" } } }, "required": true }, "responses": { "200": { "description": "Embeddings", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/EmbedResponse" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/EmbedResponse" } } } }, "400": { "description": "Batch is empty", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Batch is empty", "error_type": "empty" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } }, "413": { "description": "Batch size error", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Batch size error", "error_type": "validation" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } }, "422": { "description": "Tokenization error", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Tokenization error", "error_type": "tokenizer" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } }, "424": { "description": "Embedding Error", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Inference failed", "error_type": "backend" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } }, "429": { "description": "Model is overloaded", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Model is overloaded", "error_type": "overloaded" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } } } } }, "/embed_all": { "post": { "tags": [ "infinity_v3" ], "summary": "Get all Embeddings without Pooling.", "description": "Returns a 424 status code if the model is not an embedding model.", "operationId": "embed_all", "requestBody": { "content": { "application/json": { "schema": { "$ref": "#/components/schemas/EmbedAllRequest" } } }, "required": true }, "responses": { "200": { "description": "Embeddings", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/EmbedAllResponse" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/EmbedAllResponse" } } } }, "400": { "description": "Batch is empty", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Batch is empty", "error_type": "empty" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } }, "413": { "description": "Batch size error", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Batch size error", "error_type": "validation" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } }, "422": { "description": "Tokenization error", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Tokenization error", "error_type": "tokenizer" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } }, "424": { "description": "Embedding Error", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Inference failed", "error_type": "backend" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } }, "429": { "description": "Model is overloaded", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Model is overloaded", "error_type": "overloaded" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } } } } }, "/embed_sparse": { "post": { "tags": [ "infinity_v3" ], "summary": "Get Sparse Embeddings. Returns a 424 status code if the model is not an embedding model with SPLADE pooling.", "operationId": "embed_sparse", "requestBody": { "content": { "application/json": { "schema": { "$ref": "#/components/schemas/EmbedSparseRequest" } } }, "required": true }, "responses": { "200": { "description": "Embeddings", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/EmbedSparseResponse" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/EmbedSparseResponse" } } } }, "400": { "description": "Batch is empty", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Batch is empty", "error_type": "empty" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } }, "413": { "description": "Batch size error", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Batch size error", "error_type": "validation" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } }, "422": { "description": "Tokenization error", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Tokenization error", "error_type": "tokenizer" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } }, "424": { "description": "Embedding Error", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Inference failed", "error_type": "backend" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } }, "429": { "description": "Model is overloaded", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Model is overloaded", "error_type": "overloaded" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } } } } }, "/health": { "get": { "tags": [ "infinity_v3" ], "summary": "Health check method", "operationId": "health", "responses": { "200": { "description": "Everything is working fine", "content": { "application/json": { "schema": { "nullable": true, "enum": [ null ] } }, "application/msgpack": { "schema": { "nullable": true, "enum": [ null ] } } } }, "500": { "description": "One or more deployed models are unhealthy", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "unhealthy", "error_type": "unhealthy" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } } } } }, "/info": { "get": { "tags": [ "infinity_v3" ], "summary": "infinity_v3 endpoint info", "operationId": "get_model_info", "responses": { "200": { "description": "Served model info", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/Info" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/Info" } } } }, "400": { "description": "Missing or unknown model", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } } }, "parameters": [ { "name": "model", "in": "query", "description": "Served model name; required when multiple models are deployed", "required": false, "schema": { "type": "string", "nullable": true } } ] } }, "/metrics": { "get": { "tags": [ "infinity_v3" ], "summary": "Prometheus metrics scrape endpoint", "operationId": "metrics", "responses": { "200": { "description": "Prometheus Metrics", "content": { "text/plain": { "schema": { "type": "string" } }, "application/msgpack": { "schema": { "type": "string" } }, "application/json": { "schema": { "type": "string" } } } } } } }, "/predict": { "post": { "tags": [ "infinity_v3" ], "summary": "Deprecated: use /v1/classify. This route retains the legacy label/score response.", "operationId": "predict", "requestBody": { "content": { "application/json": { "schema": { "$ref": "#/components/schemas/PredictRequest" } } }, "required": true }, "responses": { "200": { "description": "Predictions", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/PredictResponse" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/PredictResponse" } } } }, "400": { "description": "Batch is empty", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Batch is empty", "error_type": "empty" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } }, "413": { "description": "Batch size error", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Batch size error", "error_type": "validation" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } }, "422": { "description": "Tokenization error", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Tokenization error", "error_type": "tokenizer" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } }, "424": { "description": "Prediction Error", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Inference failed", "error_type": "backend" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } }, "429": { "description": "Model is overloaded", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Model is overloaded", "error_type": "overloaded" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } } }, "deprecated": true } }, "/rerank": { "post": { "tags": [ "infinity_v3" ], "summary": "Get Ranks. Returns a 424 status code if the model is not a Sequence Classification model with", "description": "a single class.", "operationId": "rerank", "requestBody": { "content": { "application/json": { "schema": { "$ref": "#/components/schemas/RerankRequest" } } }, "required": true }, "responses": { "200": { "description": "Ranks", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/RerankResponse" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/RerankResponse" } } } }, "400": { "description": "Batch is empty", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Batch is empty", "error_type": "empty" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } }, "413": { "description": "Batch size error", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Batch size error", "error_type": "validation" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } }, "422": { "description": "Tokenization error", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Tokenization error", "error_type": "tokenizer" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } }, "424": { "description": "Rerank Error", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Inference failed", "error_type": "backend" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } }, "429": { "description": "Model is overloaded", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Model is overloaded", "error_type": "overloaded" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } } } } }, "/similarity": { "post": { "tags": [ "infinity_v3" ], "summary": "Get Sentence Similarity. Returns a 424 status code if the model is not an embedding model.", "operationId": "similarity", "requestBody": { "content": { "application/json": { "schema": { "$ref": "#/components/schemas/SimilarityRequest" } } }, "required": true }, "responses": { "200": { "description": "Sentence Similarity", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/SimilarityResponse" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/SimilarityResponse" } } } }, "400": { "description": "Batch is empty", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Batch is empty", "error_type": "empty" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } }, "413": { "description": "Batch size error", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Batch size error", "error_type": "validation" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } }, "422": { "description": "Tokenization error", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Tokenization error", "error_type": "tokenizer" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } }, "424": { "description": "Embedding Error", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Inference failed", "error_type": "backend" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } }, "429": { "description": "Model is overloaded", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Model is overloaded", "error_type": "overloaded" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } } } } }, "/tokenize": { "post": { "tags": [ "infinity_v3" ], "summary": "Tokenize inputs", "operationId": "tokenize", "requestBody": { "content": { "application/json": { "schema": { "$ref": "#/components/schemas/TokenizeRequest" } } }, "required": true }, "responses": { "200": { "description": "Tokenized ids", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/TokenizeResponse" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/TokenizeResponse" } } } }, "400": { "description": "Batch is empty", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Batch is empty", "error_type": "empty" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } }, "413": { "description": "Batch size error", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Batch size error", "error_type": "validation" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } }, "422": { "description": "Tokenization error", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" }, "example": { "error": "Tokenization error", "error_type": "tokenizer" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } } } } }, "/v1/embeddings": { "post": { "tags": [ "infinity_v3" ], "summary": "OpenAI compatible route. Returns a 424 status code if the model is not an embedding model.", "operationId": "openai_embed", "requestBody": { "content": { "application/json": { "schema": { "$ref": "#/components/schemas/OpenAICompatRequest" } } }, "required": true }, "responses": { "200": { "description": "Embeddings", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/OpenAICompatResponse" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/OpenAICompatResponse" } } } }, "400": { "description": "Batch is empty", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/OpenAICompatErrorResponse" }, "example": { "message": "Batch is empty", "type": "empty" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/OpenAICompatErrorResponse" } } } }, "413": { "description": "Batch size error", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/OpenAICompatErrorResponse" }, "example": { "message": "Batch size error", "type": "validation" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/OpenAICompatErrorResponse" } } } }, "422": { "description": "Tokenization error", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/OpenAICompatErrorResponse" }, "example": { "message": "Tokenization error", "type": "tokenizer" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/OpenAICompatErrorResponse" } } } }, "424": { "description": "Embedding Error", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/OpenAICompatErrorResponse" }, "example": { "message": "Inference failed", "type": "backend" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/OpenAICompatErrorResponse" } } } }, "429": { "description": "Model is overloaded", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/OpenAICompatErrorResponse" }, "example": { "message": "Model is overloaded", "type": "overloaded" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/OpenAICompatErrorResponse" } } } } } } }, "/v1/classify": { "post": { "tags": [ "infinity_v3" ], "operationId": "classify", "requestBody": { "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ClassifyRequest" } } }, "required": true }, "responses": { "200": { "description": "Classification probabilities and token usage", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ClassifyResponse" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ClassifyResponse" } } } }, "400": { "description": "Invalid classification input", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ClassifyErrorResponse" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ClassifyErrorResponse" } } } }, "413": { "description": "Batch or token limit exceeded", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ClassifyErrorResponse" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ClassifyErrorResponse" } } } }, "422": { "description": "Invalid JSON or tokenization", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ClassifyErrorResponse" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ClassifyErrorResponse" } } } }, "424": { "description": "Classification backend failure", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ClassifyErrorResponse" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ClassifyErrorResponse" } } } }, "429": { "description": "Model is overloaded", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ClassifyErrorResponse" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ClassifyErrorResponse" } } } } } } }, "/v1/models": { "get": { "tags": [ "infinity_v3" ], "operationId": "models", "responses": { "200": { "description": "Resident models and capabilities", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ServedModels" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ServedModels" } } } } } } }, "/v1/systemone": { "post": { "tags": [ "super::systemone" ], "operationId": "systemone", "requestBody": { "content": { "application/json": { "schema": { "$ref": "#/components/schemas/SystemOneRequestBody" }, "examples": { "joint_image": { "summary": "Perplexity joint v1.2: shared top-level image", "value": { "model": "perplexity-ai/pplx-decider-flash-v1.2", "state": "Inspect this delivery.", "images": [ "https://bucket.s3.amazonaws.com/photo.jpg" ], "questions": { "damaged": { "type": "noul", "instructions": "Is the package damaged?" } } } }, "onejev_image": { "summary": "OneJev: image inside state.messages", "value": { "model": "decision-models/baseten-elephant-9b-latest", "state": { "messages": [ { "role": "user", "content": [ { "type": "text", "text": "Inspect this delivery." }, { "type": "image_url", "image_url": { "url": "https://bucket.s3.amazonaws.com/photo.jpg" } } ] } ] }, "questions": { "damaged": { "type": "noul", "instructions": "Is the package damaged?" } } } } } } }, "required": true }, "responses": { "200": { "description": "Typed decisions", "content": { "application/json": { "schema": {} }, "application/msgpack": { "schema": {} } } }, "400": { "description": "Invalid or unsupported model", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } }, "422": { "description": "Invalid request for the selected protocol, question or token budget", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } }, "429": { "description": "Server overloaded", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } }, "application/msgpack": { "schema": { "$ref": "#/components/schemas/ErrorResponse" } } } } } } } }, "components": { "schemas": { "ClassifierModel": { "type": "object", "required": [ "id2label", "label2id" ], "properties": { "id2label": { "type": "object", "additionalProperties": { "type": "string" }, "example": { "0": "LABEL" } }, "label2id": { "type": "object", "additionalProperties": { "type": "integer", "minimum": 0 }, "example": { "LABEL": 0 } } } }, "DecodeRequest": { "type": "object", "required": [ "ids" ], "properties": { "ids": { "$ref": "#/components/schemas/InputIds" }, "skip_special_tokens": { "type": "boolean", "default": "true", "example": "true" }, "model": { "type": "string", "description": "Served model name; required for multi-model deployments.", "nullable": true } } }, "DecodeResponse": { "type": "array", "items": { "type": "string" }, "example": [ "test" ] }, "EmbedAllRequest": { "type": "object", "required": [ "inputs" ], "properties": { "inputs": { "$ref": "#/components/schemas/EmbeddingInput" }, "prompt_name": { "type": "string", "description": "The name of the prompt that should be used by for encoding. If not set, no prompt\nwill be applied.\n\nMust be a key in the `sentence-transformers` configuration `prompts` dictionary.\n\nFor example if ``prompt_name`` is \"query\" and the ``prompts`` is {\"query\": \"query: \", ...},\nthen the sentence \"What is the capital of France?\" will be encoded as\n\"query: What is the capital of France?\" because the prompt text will be prepended before\nany text to encode.", "default": "null", "example": "null", "nullable": true }, "truncate": { "type": "boolean", "default": "false", "example": "false", "nullable": true }, "truncation_direction": { "allOf": [ { "$ref": "#/components/schemas/TruncationDirection" } ], "default": "right" }, "model": { "type": "string", "description": "Served model name; required for multi-model deployments.", "nullable": true }, "encoding_format": { "allOf": [ { "$ref": "#/components/schemas/EncodingFormat" } ], "default": "float" } } }, "EmbedAllResponse": { "type": "array", "items": { "$ref": "#/components/schemas/AllEmbedding" }, "example": [ [ [ 0.0, 1.0, 2.0 ] ] ] }, "EmbedRequest": { "type": "object", "required": [ "inputs" ], "properties": { "dimensions": { "type": "integer", "description": "The number of dimensions that the output embeddings should have. If not set, the original\nshape of the representation will be returned instead.", "default": "null", "example": "null", "nullable": true, "minimum": 0 }, "inputs": { "$ref": "#/components/schemas/EmbeddingInput" }, "normalize": { "type": "boolean", "default": "true", "example": "true" }, "prompt_name": { "type": "string", "description": "The name of the prompt that should be used by for encoding. If not set, no prompt\nwill be applied.\n\nMust be a key in the `sentence-transformers` configuration `prompts` dictionary.\n\nFor example if ``prompt_name`` is \"query\" and the ``prompts`` is {\"query\": \"query: \", ...},\nthen the sentence \"What is the capital of France?\" will be encoded as\n\"query: What is the capital of France?\" because the prompt text will be prepended before\nany text to encode.", "default": "null", "example": "null", "nullable": true }, "truncate": { "type": "boolean", "default": "false", "example": "false", "nullable": true }, "truncation_direction": { "allOf": [ { "$ref": "#/components/schemas/TruncationDirection" } ], "default": "right" }, "model": { "type": "string", "description": "Served model name; required for multi-model deployments.", "nullable": true }, "encoding_format": { "allOf": [ { "$ref": "#/components/schemas/EncodingFormat" } ], "default": "float" } } }, "EmbedResponse": { "type": "array", "items": { "$ref": "#/components/schemas/Embedding" }, "example": [ [ 0.0, 1.0, 2.0 ] ] }, "EmbedSparseRequest": { "type": "object", "required": [ "inputs" ], "properties": { "inputs": { "$ref": "#/components/schemas/EmbeddingInput" }, "prompt_name": { "type": "string", "description": "The name of the prompt that should be used by for encoding. If not set, no prompt\nwill be applied.\n\nMust be a key in the `sentence-transformers` configuration `prompts` dictionary.\n\nFor example if ``prompt_name`` is \"query\" and the ``prompts`` is {\"query\": \"query: \", ...},\nthen the sentence \"What is the capital of France?\" will be encoded as\n\"query: What is the capital of France?\" because the prompt text will be prepended before\nany text to encode.", "default": "null", "example": "null", "nullable": true }, "truncate": { "type": "boolean", "default": "false", "example": "false", "nullable": true }, "truncation_direction": { "allOf": [ { "$ref": "#/components/schemas/TruncationDirection" } ], "default": "right" }, "model": { "type": "string", "description": "Served model name; required for multi-model deployments.", "nullable": true } } }, "EmbedSparseResponse": { "type": "array", "items": { "type": "array", "items": { "$ref": "#/components/schemas/SparseValue" } } }, "Embedding": { "oneOf": [ { "type": "array", "items": { "type": "number", "format": "float" } }, { "$ref": "#/components/schemas/AllEmbedding" }, { "type": "string" } ] }, "EmbeddingModel": { "type": "object", "required": [ "pooling" ], "properties": { "pooling": { "type": "string", "example": "cls" } } }, "EncodingFormat": { "type": "string", "enum": [ "float", "base64", "tensor" ] }, "ErrorResponse": { "type": "object", "required": [ "error", "error_type" ], "properties": { "error": { "type": "string" }, "error_type": { "$ref": "#/components/schemas/ErrorType" } } }, "ErrorType": { "type": "string", "enum": [ "Unhealthy", "Backend", "Overloaded", "Validation", "Tokenizer", "Empty" ] }, "Info": { "type": "object", "required": [ "model_id", "model_dtype", "model_type", "max_concurrent_requests", "max_input_length", "max_batch_tokens", "max_client_batch_size", "max_decision_questions", "auto_truncate", "tokenization_workers", "tokenization_queue_capacity", "radix_mlp_threshold", "enable_fp8_dynamic", "version" ], "properties": { "auto_truncate": { "type": "boolean" }, "docker_label": { "type": "string", "example": "null", "nullable": true }, "max_batch_requests": { "type": "integer", "default": "null", "example": "null", "nullable": true, "minimum": 0 }, "max_batch_tokens": { "type": "integer", "example": "2048", "minimum": 0 }, "max_client_batch_size": { "type": "integer", "example": "32", "minimum": 0 }, "max_concurrent_requests": { "type": "integer", "description": "Router Parameters", "example": "128", "minimum": 0 }, "max_input_length": { "type": "integer", "example": "512", "minimum": 0 }, "model_dtype": { "type": "string", "example": "float16" }, "model_id": { "type": "string", "description": "Model info", "example": "thenlper/gte-base" }, "model_sha": { "type": "string", "example": "fca14538aa9956a46526bd1d0d11d69e19b5a101", "nullable": true }, "model_type": { "$ref": "#/components/schemas/ModelType" }, "sha": { "type": "string", "example": "null", "nullable": true }, "tokenization_queue_capacity": { "type": "integer", "description": "Maximum waiting tokenizer jobs (individual inputs, not HTTP requests).", "example": "1024", "minimum": 0 }, "tokenization_workers": { "type": "integer", "example": "4", "minimum": 0 }, "version": { "type": "string", "description": "Router Info", "example": "0.5.0" }, "enable_fp8_dynamic": { "type": "boolean", "description": "Experimental per-token activation and per-row weight FP8 MLP quantization." }, "max_decision_questions": { "type": "integer", "description": "Maximum question fanout for /v1/systemone, independent of the client batch limit.", "minimum": 0 }, "radix_mlp_threshold": { "type": "number", "format": "float", "example": "0.5" } } }, "InputIds": { "oneOf": [ { "type": "array", "items": { "type": "integer", "format": "int32", "minimum": 0 } }, { "type": "array", "items": { "type": "array", "items": { "type": "integer", "format": "int32", "minimum": 0 } } } ] }, "ModelType": { "oneOf": [ { "type": "string", "enum": [ "decision" ] }, { "type": "object", "required": [ "classifier" ], "properties": { "classifier": { "$ref": "#/components/schemas/ClassifierModel" } } }, { "type": "object", "required": [ "embedding" ], "properties": { "embedding": { "$ref": "#/components/schemas/EmbeddingModel" } } }, { "type": "object", "required": [ "reranker" ], "properties": { "reranker": { "$ref": "#/components/schemas/ClassifierModel" } } } ] }, "OpenAICompatEmbedding": { "type": "object", "required": [ "object", "embedding", "index" ], "properties": { "embedding": { "$ref": "#/components/schemas/Embedding" }, "index": { "type": "integer", "example": "0", "minimum": 0 }, "object": { "type": "string", "example": "embedding" } } }, "OpenAICompatErrorResponse": { "type": "object", "required": [ "message", "code", "error_type" ], "properties": { "code": { "type": "integer", "format": "int32", "minimum": 0 }, "error_type": { "$ref": "#/components/schemas/ErrorType" }, "message": { "type": "string" } } }, "OpenAICompatRequest": { "oneOf": [ { "type": "object", "required": [ "input" ], "properties": { "dimensions": { "type": "integer", "default": "null", "example": "null", "nullable": true, "minimum": 0 }, "encoding_format": { "allOf": [ { "$ref": "#/components/schemas/EncodingFormat" } ], "default": "float" }, "input": { "$ref": "#/components/schemas/EmbeddingInput" }, "messages": { "type": "array", "items": { "$ref": "#/components/schemas/Message" }, "minItems": 1 }, "model": { "type": "string", "example": "null", "nullable": true }, "user": { "type": "string", "example": "null", "nullable": true } } }, { "type": "object", "required": [ "messages" ], "properties": { "dimensions": { "type": "integer", "default": "null", "example": "null", "nullable": true, "minimum": 0 }, "encoding_format": { "allOf": [ { "$ref": "#/components/schemas/EncodingFormat" } ], "default": "float" }, "input": { "$ref": "#/components/schemas/EmbeddingInput" }, "messages": { "type": "array", "items": { "$ref": "#/components/schemas/Message" }, "minItems": 1 }, "model": { "type": "string", "example": "null", "nullable": true }, "user": { "type": "string", "example": "null", "nullable": true } } } ], "description": "Provide exactly one of input or messages; a message list is one conversation." }, "OpenAICompatResponse": { "type": "object", "required": [ "object", "data", "model", "usage" ], "properties": { "data": { "type": "array", "items": { "$ref": "#/components/schemas/OpenAICompatEmbedding" } }, "model": { "type": "string", "example": "thenlper/gte-base" }, "object": { "type": "string", "example": "list" }, "usage": { "$ref": "#/components/schemas/OpenAICompatUsage" } } }, "OpenAICompatUsage": { "type": "object", "required": [ "prompt_tokens", "total_tokens" ], "properties": { "prompt_tokens": { "type": "integer", "example": "512", "minimum": 0 }, "total_tokens": { "type": "integer", "example": "512", "minimum": 0 } } }, "PredictInput": { "oneOf": [ { "type": "string", "description": "A single string" }, { "type": "array", "items": { "type": "string" }, "description": "A pair of strings", "maxItems": 2, "minItems": 2 }, { "type": "array", "items": { "oneOf": [ { "type": "array", "items": { "type": "string" }, "description": "A single string", "maxItems": 1, "minItems": 1 }, { "type": "array", "items": { "type": "string" }, "description": "A pair of strings", "maxItems": 2, "minItems": 2 } ] }, "description": "A batch" }, { "$ref": "#/components/schemas/MessageInput" }, { "type": "array", "items": { "$ref": "#/components/schemas/MessageInput" }, "description": "A batch of independent conversations", "minItems": 1 }, { "type": "array", "items": { "type": "integer", "maximum": 4294967295, "minimum": 0 }, "description": "A final, unpadded token-ID sequence; segment IDs are zero", "minItems": 1 }, { "type": "array", "items": { "type": "array", "items": { "type": "integer", "maximum": 4294967295, "minimum": 0 }, "description": "A final, unpadded token-ID sequence; segment IDs are zero", "minItems": 1 }, "description": "A batch of independent final token-ID sequences", "minItems": 1 } ], "description": "Model input: a string, a string pair, a batch of single strings and pairs, conversations, or final token IDs and batches of final token IDs.", "example": "What is Deep Learning?" }, "PredictRequest": { "type": "object", "required": [ "inputs" ], "properties": { "inputs": { "$ref": "#/components/schemas/PredictInput" }, "raw_scores": { "type": "boolean", "default": "false", "example": "false" }, "truncate": { "type": "boolean", "default": "false", "example": "false", "nullable": true }, "truncation_direction": { "allOf": [ { "$ref": "#/components/schemas/TruncationDirection" } ], "default": "right" }, "model": { "type": "string", "description": "Served model name; required for multi-model deployments.", "nullable": true } } }, "PredictResponse": { "oneOf": [ { "type": "array", "items": { "$ref": "#/components/schemas/Prediction" } }, { "type": "array", "items": { "type": "array", "items": { "$ref": "#/components/schemas/Prediction" } } } ] }, "Prediction": { "type": "object", "required": [ "score", "label" ], "properties": { "label": { "type": "string", "example": "admiration" }, "score": { "type": "number", "format": "float", "example": "0.5" } } }, "Rank": { "type": "object", "required": [ "index", "score" ], "properties": { "index": { "type": "integer", "example": "0", "minimum": 0 }, "score": { "type": "number", "format": "float", "example": "1.0" }, "text": { "type": "string", "default": "null", "example": "Deep Learning is ...", "nullable": true } } }, "RerankRequest": { "type": "object", "required": [ "query", "texts" ], "properties": { "query": { "type": "string", "example": "What is Deep Learning?" }, "raw_scores": { "type": "boolean", "default": "false", "example": "false" }, "return_text": { "type": "boolean", "default": "false", "example": "false" }, "texts": { "type": "array", "items": { "type": "string" }, "example": [ "Deep Learning is ..." ] }, "truncate": { "type": "boolean", "default": "false", "example": "false", "nullable": true }, "truncation_direction": { "allOf": [ { "$ref": "#/components/schemas/TruncationDirection" } ], "default": "right" }, "model": { "type": "string", "description": "Served model name; required for multi-model deployments.", "nullable": true } } }, "RerankResponse": { "type": "array", "items": { "$ref": "#/components/schemas/Rank" } }, "SimilarityInput": { "type": "object", "required": [ "source_sentence", "sentences" ], "properties": { "sentences": { "type": "array", "items": { "type": "string" }, "description": "A list of strings which will be compared against the source_sentence.", "example": [ "What is Machine Learning?" ] }, "source_sentence": { "type": "string", "description": "The string that you wish to compare the other strings with. This can be a phrase, sentence,\nor longer passage, depending on the model being used.", "example": "What is Deep Learning?" } } }, "SimilarityParameters": { "type": "object", "properties": { "prompt_name": { "type": "string", "description": "The name of the prompt that should be used by for encoding. If not set, no prompt\nwill be applied.\n\nMust be a key in the `sentence-transformers` configuration `prompts` dictionary.\n\nFor example if ``prompt_name`` is \"query\" and the ``prompts`` is {\"query\": \"query: \", ...},\nthen the sentence \"What is the capital of France?\" will be encoded as\n\"query: What is the capital of France?\" because the prompt text will be prepended before\nany text to encode.", "default": "null", "example": "null", "nullable": true }, "truncate": { "type": "boolean", "default": "false", "example": "false", "nullable": true }, "truncation_direction": { "allOf": [ { "$ref": "#/components/schemas/TruncationDirection" } ], "default": "right" } } }, "SimilarityRequest": { "type": "object", "required": [ "inputs" ], "properties": { "inputs": { "$ref": "#/components/schemas/SimilarityInput" }, "parameters": { "allOf": [ { "$ref": "#/components/schemas/SimilarityParameters" } ], "default": "null", "nullable": true }, "model": { "type": "string", "description": "Served model name; required for multi-model deployments.", "nullable": true } } }, "SimilarityResponse": { "type": "array", "items": { "type": "number", "format": "float" }, "example": [ 0.0, 1.0, 0.5 ] }, "SimpleToken": { "type": "object", "required": [ "id", "text", "special" ], "properties": { "id": { "type": "integer", "format": "int32", "example": 0, "minimum": 0 }, "special": { "type": "boolean", "example": "false" }, "start": { "type": "integer", "example": 0, "nullable": true, "minimum": 0 }, "stop": { "type": "integer", "example": 2, "nullable": true, "minimum": 0 }, "text": { "type": "string", "example": "test" } } }, "SparseValue": { "type": "object", "required": [ "index", "value" ], "properties": { "index": { "type": "integer", "minimum": 0 }, "value": { "type": "number", "format": "float" } } }, "TokenizeInput": { "oneOf": [ { "type": "string" }, { "type": "array", "items": { "type": "string" } } ] }, "TokenizeRequest": { "type": "object", "required": [ "inputs" ], "properties": { "add_special_tokens": { "type": "boolean", "default": "true", "example": "true" }, "inputs": { "$ref": "#/components/schemas/TokenizeInput" }, "prompt_name": { "type": "string", "description": "The name of the prompt that should be used by for encoding. If not set, no prompt\nwill be applied.\n\nMust be a key in the `sentence-transformers` configuration `prompts` dictionary.\n\nFor example if ``prompt_name`` is \"query\" and the ``prompts`` is {\"query\": \"query: \", ...},\nthen the sentence \"What is the capital of France?\" will be encoded as\n\"query: What is the capital of France?\" because the prompt text will be prepended before\nany text to encode.", "default": "null", "example": "null", "nullable": true }, "model": { "type": "string", "description": "Served model name; required for multi-model deployments.", "nullable": true } } }, "TokenizeResponse": { "type": "array", "items": { "type": "array", "items": { "$ref": "#/components/schemas/SimpleToken" } }, "example": [ [ { "id": 0, "special": false, "start": 0, "stop": 2, "text": "test" } ] ] }, "TruncationDirection": { "type": "string", "enum": [ "Left", "Right" ] }, "AudioFormat": { "type": "string", "enum": [ "wav", "mp3" ] }, "AudioSource": { "type": "object", "required": [ "data", "format" ], "properties": { "data": { "type": "string", "description": "Base64 encoded bytes. Not decoded by request deserialization." }, "format": { "$ref": "#/components/schemas/AudioFormat" } }, "additionalProperties": false }, "ContentPart": { "oneOf": [ { "type": "object", "required": [ "text", "type" ], "properties": { "text": { "type": "string" }, "type": { "type": "string", "enum": [ "text" ] } } }, { "type": "object", "required": [ "image_url", "type" ], "properties": { "image_url": { "$ref": "#/components/schemas/ImageSource" }, "type": { "type": "string", "enum": [ "image_url" ] } } }, { "type": "object", "required": [ "input_audio", "type" ], "properties": { "input_audio": { "$ref": "#/components/schemas/AudioSource" }, "type": { "type": "string", "enum": [ "input_audio" ] } } }, { "type": "object", "required": [ "video_url", "type" ], "properties": { "type": { "type": "string", "enum": [ "video_url" ] }, "video_url": { "$ref": "#/components/schemas/VideoSource" } } } ], "description": "Ordered wire content, independent of a particular model's tensor format.\nThese types describe inputs; the loaded adapter must explicitly support each\nmodality. VideoUrl is an extension following the vLLM message convention.", "discriminator": { "propertyName": "type" } }, "EmbeddingInput": { "oneOf": [ { "type": "string" }, { "type": "array", "items": { "type": "string" } }, { "type": "array", "items": { "type": "integer", "format": "int32", "minimum": 0 } }, { "type": "array", "items": { "type": "array", "items": { "type": "integer", "format": "int32", "minimum": 0 } } }, { "type": "array", "items": { "$ref": "#/components/schemas/Message" } } ], "description": "Embedding input: text, final token IDs, batches of either, or one conversation.\nA message list always describes one input, regardless of the number of turns.\nAn empty array is parsed as an empty text batch and rejected by the request handler." }, "ImageDetail": { "type": "string", "enum": [ "auto", "low", "high" ] }, "ImageSource": { "type": "object", "required": [ "url" ], "properties": { "detail": { "allOf": [ { "$ref": "#/components/schemas/ImageDetail" } ], "nullable": true }, "url": { "type": "string", "description": "Remote URL or inline data URL; fetching is a processor responsibility." } }, "additionalProperties": false }, "Message": { "type": "object", "required": [ "role", "content" ], "properties": { "content": { "$ref": "#/components/schemas/MessageContent" }, "role": { "$ref": "#/components/schemas/MessageRole" } }, "additionalProperties": false }, "MessageContent": { "oneOf": [ { "type": "string" }, { "type": "array", "items": { "$ref": "#/components/schemas/ContentPart" } } ] }, "MessageRole": { "type": "string", "description": "Roles supported by the initial conversation contract. Additional roles require\nexplicit metadata and adapter support, rather than being silently discarded.", "enum": [ "system", "developer", "user", "assistant", "query", "document" ] }, "VideoSource": { "type": "object", "required": [ "url" ], "properties": { "url": { "type": "string" } }, "additionalProperties": false }, "MessageInput": { "type": "object", "required": [ "messages" ], "properties": { "messages": { "type": "array", "items": { "$ref": "#/components/schemas/Message" } } }, "additionalProperties": false }, "ClassifyRequest": { "allOf": [ { "type": "object", "description": "SGLang-compatible classification, with native messages and truncation extensions.\nProvide exactly one of input or messages. String arrays are independent inputs.", "properties": { "input": { "$ref": "#/components/schemas/EmbeddingInput" }, "messages": { "type": "array", "items": { "$ref": "#/components/schemas/Message" }, "minItems": 1 }, "model": { "type": "string", "description": "Served model name; required for multi-model deployments.", "nullable": true }, "priority": { "type": "integer", "format": "int32", "description": "Only absent or zero priority is supported.", "nullable": true }, "rid": { "allOf": [ { "$ref": "#/components/schemas/ClassifyRequestId" } ], "nullable": true }, "truncate": { "type": "boolean", "nullable": true }, "truncation_direction": { "$ref": "#/components/schemas/TruncationDirection" }, "use_activation": { "type": "boolean", "description": "vLLM extension: `false` returns raw logits instead of softmax probabilities, which\none-label cross-encoders need for a usable relevance score.", "nullable": true }, "user": { "type": "string", "nullable": true } } }, { "oneOf": [ { "type": "object", "required": [ "input" ] }, { "type": "object", "required": [ "messages" ] } ] } ], "description": "Provide exactly one of input or messages; a message list is one conversation." }, "ClassifyRequestId": { "oneOf": [ { "type": "string" }, { "type": "array", "items": { "type": "string" } } ] }, "ClassifyResponse": { "type": "object", "required": [ "id", "object", "created", "model", "data", "usage" ], "properties": { "created": { "type": "integer", "format": "int64", "minimum": 0 }, "data": { "type": "array", "items": { "$ref": "#/components/schemas/ClassifyData" } }, "id": { "type": "string" }, "model": { "type": "string" }, "object": { "type": "string" }, "usage": { "$ref": "#/components/schemas/ClassifyUsage" } } }, "ClassifyData": { "type": "object", "required": [ "index", "label", "probs", "num_classes" ], "properties": { "index": { "type": "integer", "minimum": 0 }, "label": { "type": "string" }, "num_classes": { "type": "integer", "minimum": 0 }, "probs": { "type": "array", "items": { "type": "number", "format": "float" } } } }, "ClassifyUsage": { "type": "object", "required": [ "prompt_tokens", "completion_tokens", "total_tokens" ], "properties": { "completion_tokens": { "type": "integer", "minimum": 0 }, "prompt_tokens": { "type": "integer", "minimum": 0 }, "prompt_tokens_details": { "nullable": true }, "total_tokens": { "type": "integer", "minimum": 0 } } }, "ClassifyErrorResponse": { "type": "object", "required": [ "object", "message", "type", "code" ], "properties": { "code": { "type": "integer", "format": "int32", "minimum": 0 }, "message": { "type": "string" }, "object": { "type": "string" }, "param": { "type": "string", "nullable": true }, "type": { "type": "string" } } }, "ModelInput": { "oneOf": [ { "type": "string" }, { "$ref": "#/components/schemas/MessageInput" } ], "description": "A plain text state or an explicitly structured native conversation.\nMessage objects are never interpreted as text or silently flattened." }, "ServedModel": { "type": "object", "required": [ "id", "object", "checkpoint_id", "backend", "capabilities" ], "properties": { "backend": { "type": "string" }, "capabilities": { "type": "array", "items": { "type": "string" } }, "checkpoint_id": { "type": "string" }, "id": { "type": "string" }, "object": { "type": "string" } } }, "ServedModels": { "type": "object", "required": [ "object", "data" ], "properties": { "data": { "type": "array", "items": { "$ref": "#/components/schemas/ServedModel" } }, "object": { "type": "string" } } }, "SystemOneRequest": { "type": "object", "description": "OneJev (including Elephant/Mammoth), Rune, Clef, Laya and Perplexity v1.1.\nTop-level images/videos are rejected with HTTP 422. Image-capable OneJev/Rune\nmodels accept user image_url content parts inside state.messages instead.", "required": [ "state", "questions" ], "properties": { "head_max_len": { "type": "integer", "nullable": true, "minimum": 0 }, "max_len": { "type": "integer", "nullable": true, "minimum": 0 }, "model": { "type": "string", "nullable": true }, "questions": { "type": "object", "additionalProperties": {} }, "state": { "$ref": "#/components/schemas/SystemOneState" } }, "additionalProperties": false }, "SystemOneState": { "oneOf": [ { "type": "string" }, { "oneOf": [ { "type": "object", "additionalProperties": true }, { "type": "object", "required": [ "messages" ], "additionalProperties": true } ], "description": "Arbitrary JSON object without the reserved messages key." }, { "type": "array", "items": {} }, { "$ref": "#/components/schemas/MessageInput" } ], "description": "Text, a JSON array/object, or Infinity's native conversation envelope. Objects with messages must match MessageInput. Supported roles and media depend on the selected model: OneJev accepts user/assistant text and user image_url parts, with image detail omitted or auto; Laya rejects conversations. Content parts use Chat Completions-style names; this is not an OpenAI Chat Completions, Responses, or Decisions endpoint." }, "Tensor": { "type": "object", "required": [ "shape", "dtype", "data_encoding", "data" ], "properties": { "data": { "type": "string", "format": "byte", "description": "Contiguous row-major data, little-endian for multi-byte elements." }, "data_encoding": { "type": "string", "description": "\"base64\" for JSON, \"binary\" for MessagePack." }, "dtype": { "type": "string", "description": "Element dtype; currently float32.", "example": "float32" }, "shape": { "type": "array", "items": { "type": "integer", "minimum": 0 } } } }, "AllEmbedding": { "oneOf": [ { "type": "array", "items": { "type": "array", "items": { "type": "number", "format": "float" } } }, { "type": "array", "items": { "type": "string" } }, { "$ref": "#/components/schemas/Tensor" } ] }, "SystemOneRequestBody": { "anyOf": [ { "$ref": "#/components/schemas/SystemOneRequest" }, { "$ref": "#/components/schemas/D3SystemOneRequest" }, { "$ref": "#/components/schemas/PplxJointSystemOneRequest" } ], "description": "Use the schema for the selected model's decision protocol, not an arbitrary alternative. OneJev (Elephant/Mammoth), Rune, Clef, Laya and Perplexity v1.1 use SystemOneRequest. D3 and Perplexity joint v1.2 use their respective top-level media formats. Model is required when multiple models are loaded." }, "D3SystemOneRequest": { "type": "object", "description": "D3 accepts arbitrary JSON states and top-level media shared by every question.\nObjects containing messages are serialized as JSON, not interpreted as native conversations.", "required": [ "state", "questions" ], "properties": { "head_max_len": { "type": "integer", "nullable": true, "minimum": 0 }, "images": { "type": "array", "items": { "type": "string" }, "description": "Ordered remote URLs or inline data URLs, subject to configured media limits and host policy." }, "max_len": { "type": "integer", "nullable": true, "minimum": 0 }, "model": { "type": "string", "nullable": true }, "questions": { "type": "object", "additionalProperties": {} }, "state": {}, "videos": { "type": "array", "items": { "type": "string" }, "description": "Ordered video sources; supported only by the D3 protocol." } }, "additionalProperties": false }, "PplxJointSystemOneRequest": { "type": "object", "description": "Perplexity joint v1.2 (pplx-decider-flash-v1.2): arbitrary JSON state and shared top-level images. Nonempty videos are rejected. Native state.messages is serialized as JSON, not interpreted as a conversation.", "required": [ "state", "questions" ], "properties": { "head_max_len": { "type": "integer", "nullable": true, "minimum": 0 }, "images": { "type": "array", "items": { "type": "string" }, "description": "Ordered remote URLs or inline data URLs, subject to configured media limits and host policy." }, "max_len": { "type": "integer", "nullable": true, "minimum": 0 }, "model": { "type": "string", "nullable": true }, "questions": { "type": "object", "additionalProperties": {} }, "state": {}, "videos": { "type": "array", "items": { "type": "string" }, "description": "Must be omitted or empty for this protocol.", "maxItems": 0 } }, "additionalProperties": false } } }, "tags": [ { "name": "infinity_v3", "description": "infinity_v3 API" } ] }