> ## Documentation Index > Fetch the complete documentation index at: https://www.truefoundry.com/llms.txt > Use this file to discover all available pages before exploring further. # Chat Completions > Generate chat-based completions using the specified model. ## OpenAPI ````yaml /gateway-openapi.json post /chat/completions openapi: 3.0.0 info: title: LLM Gateway description: API for LLM Gateway version: 1.0.0 servers: - url: https://{gatewayBaseURL} variables: gatewayBaseURL: default: gateway.truefoundry.ai description: Gateway base URL description: Gateway base URL security: - AuthorizationBearer: [] tags: - name: Chat - name: Agent - name: Messages - name: MCP - name: Embeddings - name: Rerank - name: Responses - name: Image - name: Audio - name: Batch - name: Files - name: Fine-tuning - name: Moderations - name: Models paths: /chat/completions: post: tags: - Chat summary: Chat Completions description: Generate chat-based completions using the specified model. parameters: - schema: type: string description: Optional metadata for the request required: false name: x-tfy-metadata in: header - schema: type: string description: JSON-encoded logging configuration for the request required: false name: x-tfy-logging-config in: header requestBody: required: true description: Parameters for generating chat completions. content: application/json: schema: type: object properties: model: type: string description: Identifier for the language model to be used for generation. messages: type: array items: oneOf: - type: object properties: role: type: string enum: - system description: >- System message providing context or instructions to the model. content: anyOf: - type: string - type: array items: type: object properties: type: type: string description: >- The type of content being transmitted. Supported values include "text" and "image". text: type: string nullable: true description: >- The text content when type is "text". UTF-8 encoded string. image_url: type: object nullable: true properties: url: type: string description: >- The URL pointing to the image resource. Must be a valid, accessible URL. required: - url description: >- Image URL configuration when type is "image". required: - type description: >- System instructions or context. Can be a string or an array. name: type: string nullable: true description: Optional identifier for the system. required: - role - content title: System Message - type: object properties: role: type: string enum: - user description: Message from the end user making the request. content: anyOf: - type: string - type: array items: type: object properties: type: type: string description: >- The type of content being transmitted. Supported values include "text" and "image". text: type: string nullable: true description: >- The text content when type is "text". UTF-8 encoded string. image_url: type: object nullable: true properties: url: type: string description: >- The URL pointing to the image resource. Must be a valid, accessible URL. required: - url description: >- Image URL configuration when type is "image". required: - type description: User input as text or structured content array. name: type: string nullable: true description: >- Optional identifier for the user sending the message. required: - role - content title: User Message - type: object properties: role: type: string enum: - assistant description: Response message from the AI assistant. content: anyOf: - type: string - type: array items: type: object properties: type: type: string description: >- The type of content being transmitted. Supported values include "text" and "image". text: type: string nullable: true description: >- The text content when type is "text". UTF-8 encoded string. image_url: type: object nullable: true properties: url: type: string description: >- The URL pointing to the image resource. Must be a valid, accessible URL. required: - url description: >- Image URL configuration when type is "image". required: - type - {} - {} description: >- Assistant's response content. Can be text, structured content, or null when using tools/functions. function_call: description: >- Function call details when the assistant invokes a function. tool_calls: description: >- Tool call details when the assistant uses available tools. name: type: string nullable: true description: Optional identifier for the assistant instance. citationMetadata: type: object nullable: true properties: citationSources: type: array nullable: true items: type: object properties: startIndex: type: number nullable: true description: >- Starting character index in the content where this citation applies. endIndex: type: number nullable: true description: >- Ending character index in the content where this citation applies. uri: type: string nullable: true description: >- URI reference to the source material being cited. license: type: string nullable: true description: >- License information for the cited content. description: >- Array of citation sources referenced in the content. description: Metadata about citations used in the response. required: - role title: Assistant Message - type: object properties: role: type: string enum: - function description: Message containing function execution results. content: type: string description: The result returned from the function execution. name: type: string description: Name of the function that was executed. required: - role - content - name title: Function Message - type: object properties: role: type: string enum: - tool description: Message containing tool execution results. content: anyOf: - type: string - type: array items: type: object properties: type: type: string description: >- The type of content being transmitted. Supported values include "text" and "image". text: type: string nullable: true description: >- The text content when type is "text". UTF-8 encoded string. image_url: type: object nullable: true properties: url: type: string description: >- The URL pointing to the image resource. Must be a valid, accessible URL. required: - url description: >- Image URL configuration when type is "image". required: - type description: >- The result returned from the tool execution. Can be a string or an array of content parts. tool_call_id: type: string description: >- Unique identifier for the tool call this message responds to. required: - role - content - tool_call_id title: Tool Message - type: object properties: role: type: string enum: - developer description: Message from a developer or system administrator. content: anyOf: - type: string - type: array items: type: object properties: type: type: string description: >- The type of content being transmitted. Supported values include "text" and "image". text: type: string nullable: true description: >- The text content when type is "text". UTF-8 encoded string. image_url: type: object nullable: true properties: url: type: string description: >- The URL pointing to the image resource. Must be a valid, accessible URL. required: - url description: >- Image URL configuration when type is "image". required: - type description: >- Developer message content or system configuration instructions. Can be a string or an array. name: type: string nullable: true description: Optional identifier for the developer. required: - role - content title: Developer Message description: >- Conversation history as an array of messages for contextual generation. tools: type: array nullable: true items: type: object properties: type: type: string description: >- The type of tool being specified. Currently supports "function". function: type: object nullable: true properties: name: type: string description: >- The name of the function to be called. Must be unique within the context. description: type: string nullable: true description: >- Detailed description of the function's purpose and behavior. parameters: type: object nullable: true additionalProperties: {} description: >- JSON Schema specification defining the expected parameters for this function. required: - name description: Function specification when type is "function". required: - type description: >- Array of tool definitions available for the model to use during generation. tool_choice: anyOf: - type: string enum: - none - type: string enum: - auto - type: string enum: - required - type: object properties: type: type: string enum: - function function: type: object properties: name: type: string description: The function the model should call. required: - name required: - type - function - {} description: >- Controls tool usage behavior. Use "none" to disable, "auto" for model choice, "required" to force tool use, or specify tool for forced usage. temperature: type: number nullable: true description: >- Sampling temperature between 0 and 2. Higher values make output more random, lower values more deterministic. top_p: type: number nullable: true description: >- Nucleus sampling parameter. Limits cumulative probability of tokens considered for sampling. top_k: type: number nullable: true description: Limits sampling to the K most-likely next tokens. 'n': type: number nullable: true description: Number of response alternatives to generate. stream: type: boolean nullable: true description: >- Enable streaming of partial response chunks as they are generated. logprobs: anyOf: - type: boolean - type: number - {} description: >- Include log probabilities of tokens: set to boolean for Chat Completions; integer for legacy Completions (number of top-token logprobs). stop: anyOf: - type: string - type: array items: type: string - {} description: >- Sequence(s) at which to stop generation. Can be a single string or array of strings. max_tokens: type: number nullable: true description: Maximum number of tokens to generate in the response. presence_penalty: type: number nullable: true description: >- Penalty factor for new tokens based on their presence in existing text. Range: -2.0 to 2.0 frequency_penalty: type: number nullable: true description: >- Penalty factor for new tokens based on their frequency in existing text. Range: -2.0 to 2.0 logit_bias: type: object nullable: true additionalProperties: type: number description: >- Modify the likelihood of specified tokens appearing in the completion. user: type: string nullable: true description: >- Unique identifier for the end-user making the request, for monitoring and detecting abuse. prompt_version_fqn: type: string nullable: true description: >- Fully qualified name (FQN) of the prompt template to use for automatic prompt hydration. prompt_variables: type: object nullable: true additionalProperties: anyOf: - type: string - type: number description: >- Key-value pairs of variables to substitute in the prompt template when using prompt_version_fqn. required: - model - messages responses: '200': description: Chat completions generated successfully. content: application/json: schema: type: object properties: id: type: string description: Id of the response. object: type: string description: Type of the object, e.g., 'chat.completion'. created: type: number description: Timestamp of when the response was created. model: type: string description: The model used to generate the response. choices: type: array items: type: object properties: index: type: number description: Index of the choice. message: type: object properties: role: type: string description: Role of the message sender, e.g., 'assistant'. content: type: string nullable: true description: Content of the message. refusal: type: string nullable: true description: Refusal of the message. required: - role - content - refusal logprobs: description: Log probabilities of the tokens. finish_reason: type: string description: Reason why the generation stopped. required: - index - message - finish_reason description: Array of choices returned by the model. usage: type: object properties: prompt_tokens: type: number description: Number of tokens in the prompt. completion_tokens: type: number description: Number of tokens in the completion. total_tokens: type: number description: Total number of tokens used. prompt_tokens_details: type: object nullable: true properties: cached_tokens: type: number description: Number of cached tokens. audio_tokens: type: number description: Number of audio tokens. required: - cached_tokens - audio_tokens completion_tokens_details: type: object nullable: true properties: reasoning_tokens: type: number description: Number of reasoning tokens. audio_tokens: type: number description: Number of audio tokens. accepted_prediction_tokens: type: number description: Number of accepted prediction tokens. rejected_prediction_tokens: type: number description: Number of rejected prediction tokens. required: - reasoning_tokens - audio_tokens - accepted_prediction_tokens - rejected_prediction_tokens required: - prompt_tokens - completion_tokens - total_tokens description: Details about token usage. service_tier: type: string nullable: true description: Service tier used for the request. system_fingerprint: type: string nullable: true description: System fingerprint, if available. required: - id - object - created - model - choices - usage text/event-stream: schema: type: string description: Server-sent events stream for chat completions '400': description: Bad request. components: securitySchemes: AuthorizationBearer: type: http scheme: bearer bearerFormat: JWT ````