LiteLLM Chat Completions API

openapi: 3.1.0
info:
  title: LiteLLM Proxy API
  description: >-
    Minimal OpenAPI 3.1 specification for the LiteLLM proxy server,
    covering chat completions, completions, embeddings, audio, images,
    moderations, fine-tuning, batches, assistants, vector stores,
    messages, OCR, rerank, and RAG endpoints.
  version: 1.0.0
x-generated-from: https://docs.litellm.ai/docs/completion
x-generated-by: claude-crawl-2026-05-08
servers:
  - url: http://localhost:4000
    description: LiteLLM proxy server
paths:
  /chat/completions:
    post:
      operationId: createChatCompletion
      summary: Create a chat completion
      responses:
        '200':
          description: Chat completion result
        default:
          description: Error response
  /completions:
    post:
      operationId: createCompletion
      summary: Create a text completion
      responses:
        '200':
          description: Text completion result
        default:
          description: Error response
  /embeddings:
    post:
      operationId: createEmbedding
      summary: Create an embedding vector
      responses:
        '200':
          description: Embedding result
        default:
          description: Error response
  /audio/transcriptions:
    post:
      operationId: createAudioTranscription
      summary: Transcribe audio to text
      responses:
        '200':
          description: Transcription result
        default:
          description: Error response
  /audio/speech:
    post:
      operationId: createAudioSpeech
      summary: Synthesize speech from text
      responses:
        '200':
          description: Audio response
        default:
          description: Error response
  /images/generations:
    post:
      operationId: createImageGeneration
      summary: Generate images
      responses:
        '200':
          description: Generated images
        default:
          description: Error response
  /images/edits:
    post:
      operationId: createImageEdit
      summary: Edit an image
      responses:
        '200':
          description: Edited image
        default:
          description: Error response
  /images/variations:
    post:
      operationId: createImageVariation
      summary: Generate image variations
      responses:
        '200':
          description: Image variations
        default:
          description: Error response
  /moderations:
    post:
      operationId: createModeration
      summary: Classify text for moderation
      responses:
        '200':
          description: Moderation result
        default:
          description: Error response
  /fine_tuning/jobs:
    post:
      operationId: createFineTuningJob
      summary: Create a fine-tuning job
      responses:
        '200':
          description: Fine-tuning job created
        default:
          description: Error response
  /batches:
    post:
      operationId: createBatch
      summary: Create a batch job
      responses:
        '200':
          description: Batch job created
        default:
          description: Error response
  /assistants:
    post:
      operationId: createAssistant
      summary: Create an assistant
      responses:
        '200':
          description: Assistant created
        default:
          description: Error response
  /vector_stores:
    post:
      operationId: createVectorStore
      summary: Create a vector store
      responses:
        '200':
          description: Vector store created
        default:
          description: Error response
  /vector_stores/search:
    post:
      operationId: searchVectorStore
      summary: Search a vector store
      responses:
        '200':
          description: Search result
        default:
          description: Error response
  /v1/messages:
    post:
      operationId: createMessage
      summary: Create a message (Anthropic-compatible)
      responses:
        '200':
          description: Message result
        default:
          description: Error response
  /v1/messages/count_tokens:
    post:
      operationId: countMessageTokens
      summary: Count tokens for a message
      responses:
        '200':
          description: Token count
        default:
          description: Error response
  /ocr:
    post:
      operationId: createOcr
      summary: Run OCR on an image
      responses:
        '200':
          description: OCR result
        default:
          description: Error response
  /rerank:
    post:
      operationId: rerankDocuments
      summary: Rerank documents against a query
      responses:
        '200':
          description: Rerank result
        default:
          description: Error response
  /rag/ingest:
    post:
      operationId: ragIngest
      summary: Ingest documents into a RAG index
      responses:
        '200':
          description: Ingestion accepted
        default:
          description: Error response
  /rag/query:
    post:
      operationId: ragQuery
      summary: Query a RAG index
      responses:
        '200':
          description: Query result
        default:
          description: Error response
components: {}