> ## Documentation Index > Fetch the complete documentation index at: https://concentrate.ai/docs/llms.txt > Use this file to discover all available pages before exploring further. # Chat Completions > OpenAI Chat Completions API compatibility endpoint for Cursor, Opencode, and other clients **Beta Feature** The Chat Completions API is currently in beta. It provides OpenAI Chat Completions API compatibility for clients like Cursor, Opencode, and other tools that use the OpenAI format. For production use, we recommend using the [Responses API](/docs/api-reference/endpoint/create-response) instead. ## Overview OpenAI Chat Completions API compatibility endpoint. Use Concentrate as a drop-in replacement for OpenAI in any tool or SDK that supports configurable base URLs. ## OpenAPI ````yaml POST /v1/chat/completions/ openapi: 3.0.3 info: title: Responses API description: Unified responses API version: 1.0.0 servers: - url: https://api.concentrate.ai security: [] paths: /v1/chat/completions/: post: tags: - chat-completions - openai description: >- OpenAI Chat Completions API compatibility endpoint for Cursor, Opencode, and other clients requestBody: required: true content: application/json: schema: type: object properties: messages: minItems: 1 type: array items: oneOf: - type: object properties: role: type: string enum: - developer - system description: The role of the author of this message. content: anyOf: - type: string - type: array items: type: object properties: type: type: string enum: - text description: The type of content. text: type: string description: The text content. prompt_cache_breakpoint: allOf: - $ref: >- #/components/schemas/PromptCacheBreakpointInput required: - type - text description: The contents of the system message. name: description: An optional name for the participant. nullable: true type: string required: - role - content title: System Message - type: object properties: role: type: string enum: - user description: The role of the author of this message. content: anyOf: - type: string - type: array items: oneOf: - type: object properties: type: type: string enum: - text description: The type of content. text: type: string description: The text content. prompt_cache_breakpoint: allOf: - $ref: >- #/components/schemas/PromptCacheBreakpointInput required: - type - text title: Text - type: object properties: type: type: string enum: - image_url description: The type of content. image_url: type: object properties: url: type: string description: The URL of the image. required: - url description: Image URL details. prompt_cache_breakpoint: allOf: - $ref: >- #/components/schemas/PromptCacheBreakpointInput required: - type - image_url title: Image URL - type: object properties: type: type: string enum: - file description: The type of content. file: type: object properties: filename: description: The filename of the file. type: string file_data: description: The base64-encoded file data. type: string file_url: description: The URL of the file. type: string file_id: description: The ID of the file. type: string description: File details. prompt_cache_breakpoint: allOf: - $ref: >- #/components/schemas/PromptCacheBreakpointInput required: - type - file title: File description: The contents of the user message. required: - role - content title: User Message - type: object properties: role: type: string enum: - assistant description: The role of the author of this message. content: description: The contents of the assistant message. nullable: true anyOf: - type: string - type: array items: oneOf: - type: object properties: type: type: string enum: - text description: The type of content. text: type: string description: The text content. prompt_cache_breakpoint: allOf: - $ref: >- #/components/schemas/PromptCacheBreakpointInput required: - type - text title: Text name: description: An optional name for the participant. nullable: true type: string refusal: description: Refusals from the model are returned as refusal. nullable: true type: string reasoning: description: The reasoning content generated by the model. nullable: true type: string reasoning_content: description: >- The reasoning content generated by the model. Included for compatibility. Use `reasoning` over `reasoning_content`. nullable: true type: string tool_calls: description: The tool calls generated by the model. nullable: true type: array items: oneOf: - type: object properties: type: type: string enum: - function description: The type of the tool call. id: type: string description: The ID of the tool call. function: type: object properties: name: description: The name of the function to call. type: string maxLength: 256 arguments: type: string description: The arguments to pass to the function. required: - name - arguments description: The function that the model called. required: - type - id - function title: Function Tool Call - type: object properties: type: type: string enum: - custom description: The type of the tool call. id: type: string description: The ID of the tool call. custom: type: object properties: input: type: string description: The input to pass to the custom tool. name: description: The name of the custom tool. type: string maxLength: 256 required: - input - name description: The custom tool that the model called. required: - type - id - custom title: Custom Tool Call function_call: nullable: true type: object properties: arguments: type: string description: The arguments passed to the function. name: description: The name of the function to call. type: string maxLength: 256 required: - arguments - name description: Deprecated in favor of tool_calls. required: - role title: Assistant Message - type: object properties: role: type: string enum: - tool description: The role of the author of this message. content: anyOf: - type: string - type: array items: oneOf: - type: object properties: type: type: string enum: - text description: The type of content. text: type: string description: The text content. prompt_cache_breakpoint: allOf: - $ref: >- #/components/schemas/PromptCacheBreakpointInput required: - type - text title: Text - type: object properties: type: type: string enum: - image_url description: The type of content. image_url: type: object properties: url: type: string description: The URL of the image. required: - url description: Image URL details. prompt_cache_breakpoint: allOf: - $ref: >- #/components/schemas/PromptCacheBreakpointInput required: - type - image_url title: Image URL - type: object properties: type: type: string enum: - file description: The type of content. file: type: object properties: filename: description: The filename of the file. type: string file_data: description: The base64-encoded file data. type: string file_url: description: The URL of the file. type: string file_id: description: The ID of the file. type: string description: File details. prompt_cache_breakpoint: allOf: - $ref: >- #/components/schemas/PromptCacheBreakpointInput required: - type - file title: File description: The contents of the tool message. tool_call_id: type: string description: >- The ID of the tool call this message is responding to. required: - role - content - tool_call_id title: Tool Message description: A list of messages comprising the conversation so far. model: description: >- Model identifier. Use /v1/models to list all available models. Supports canonical names (e.g. gpt-5.2, claude-opus-4-6), aliases, and provider-prefixed formats (e.g. openai/gpt-5.2). Use "auto" for automatic model selection. type: string frequency_penalty: description: >- Number between -2.0 and 2.0. Positive values penalize new tokens based on their existing frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim. nullable: true type: number minimum: -2 maximum: 2 function_call: description: >- Deprecated in favor of the tool_calls parameter. Will be converted to `tool_choice`. Using `tool_choice` will overwrite this. nullable: true anyOf: - type: string enum: - none - auto - type: object properties: name: type: string maxLength: 256 description: >- The name. Must contain only alphanumeric characters, underscores, dots, and hyphens, with a maximum length of 256 characters, with a minimum length of 1 character after unaccepted characters are stripped. required: - name additionalProperties: false title: Function functions: description: >- Deprecated in favor of the tools parameter. Will be converted to tools when used. nullable: true type: array items: type: object properties: name: type: string maxLength: 256 description: >- The name. Must contain only alphanumeric characters, underscores, dots, and hyphens, with a maximum length of 256 characters, with a minimum length of 1 character after unaccepted characters are stripped. description: nullable: true type: string parameters: type: object additionalProperties: {} required: - name - parameters additionalProperties: false logit_bias: description: >- Unsupported. Modify the likelihood of specified tokens appearing in the completion. nullable: true type: object additionalProperties: type: number minimum: -100 maximum: 100 logprobs: description: >- Whether to return log probabilities of the output tokens. If true, logprobs will be included in the response. nullable: true type: boolean max_completion_tokens: description: >- The maximum number of tokens that can be generated in the chat completion. The total length of input tokens and generated tokens is limited by the model's context length. nullable: true type: number minimum: 0 exclusiveMinimum: true max_tokens: description: Deprecated, use max_completion_tokens instead. nullable: true type: number minimum: 0 exclusiveMinimum: true metadata: description: >- Set of 16 key-value pairs that can be attached to an object. This can be useful for storing additional information about the object in a structured format. nullable: true type: object additionalProperties: type: string maxLength: 512 modalities: description: >- Unsupported. A list of modalities that the model supports. The model will use these modalities when processing the request. nullable: true minItems: 1 maxItems: 2 type: array items: type: string enum: - text - audio 'n': description: >- Unsupported. How many chat completion choices to generate for each input message. nullable: true type: integer minimum: 1 maximum: 128 parallel_tool_calls: description: Whether to enable parallel function calling during tool use. nullable: true type: boolean prediction: description: >- Unsupported. Specifies the model should predict a specified content, then continue generating naturally. nullable: true type: object properties: type: type: string enum: - content content: anyOf: - type: string - type: array items: type: object properties: type: type: string enum: - text description: The type of content. text: type: string description: The text content. prompt_cache_breakpoint: allOf: - $ref: >- #/components/schemas/PromptCacheBreakpointInput required: - type - text additionalProperties: false required: - type - content presence_penalty: description: >- Unsupported. Number between -2.0 and 2.0. Positive values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood to talk about new topics. nullable: true type: number minimum: -2 maximum: 2 prompt_cache_key: description: >- Optional custom identifier for managing prompt cache entries. nullable: true type: string prompt_cache_options: nullable: true allOf: - $ref: '#/components/schemas/PromptCacheOptionsInput' prompt_cache_retention: nullable: true type: string enum: - in-memory - in_memory - 24h description: >- The legacy OpenAI prompt-cache retention policy. Earlier OpenAI models use this field for in-memory or 24-hour retention. GPT-5.6 and later use prompt_cache_options.ttl instead. reasoning_effort: nullable: true type: string enum: - none - minimal - low - medium - high - xhigh - max description: >- Constrains effort on reasoning for reasoning models. Reducing reasoning effort can result in faster responses and fewer tokens used on reasoning in a response. Not all models support all reasoning levels. If your requested reasoning level isn't supported by the model, Concentrate bumps it up, then down, to the closest reasoning level. response_format: description: An object specifying the format that the model must output. nullable: true oneOf: - $ref: '#/components/schemas/TextFormatInput' - type: object properties: type: type: string enum: - json_schema description: The response format type. json_schema: type: object properties: name: description: The name of the response schema. type: string maxLength: 256 description: description: A description of the response schema. nullable: true type: string strict: description: Whether to enforce strict schema adherence. nullable: true type: boolean schema: type: object additionalProperties: {} description: The JSON schema for the response. required: - name - schema description: The JSON schema details. required: - type - json_schema title: JSON Schema Format - type: object properties: type: type: string enum: - json_object description: The response format type. required: - type title: JSON Object Format safety_identifier: description: >- Unique identifier to help detect when content may violate usage policies. nullable: true type: string seed: description: >- Unsupported. If specified, our system will make a best effort to sample deterministically, such that repeated requests with the same seed and parameters should return the same result. nullable: true type: string service_tier: description: Specifies which tier to use for processing the request. nullable: true type: string enum: - auto - default - flex - scale - priority stop: description: >- Unsupported. Up to 4 sequences where the API will stop generating further tokens. nullable: true anyOf: - type: string - type: array items: type: string store: description: >- If true, the model's response will be stored for reuse by other requests. nullable: true type: boolean stream: description: >- If true, partial message deltas will be sent. Tokens will be sent as data-only server-sent events as they become available. nullable: true type: boolean stream_options: nullable: true type: object properties: include_obfuscation: nullable: true type: boolean description: >- When true, stream obfuscation will be enabled. Stream obfuscation adds random characters to an obfuscation field on streaming delta events to normalize payload sizes as a mitigation to certain side-channel attacks. These obfuscation fields are included by default, but add a small amount of overhead to the data stream. You can set include_obfuscation to false to optimize for bandwidth if you trust the network links between your application and the API. Currently unsupported, but included for compatibility. include_usage: nullable: true type: boolean temperature: description: >- What sampling temperature to use, between 0 and 2. Higher values like 0.8 will make the output more random, while lower values like 0.2 will make it more focused and deterministic. nullable: true type: number minimum: 0 maximum: 2 tool_choice: description: Controls which (if any) function is called by the model. nullable: true anyOf: - type: string enum: - none - auto - required description: >- Controls which (if any) tool is called by the model. none means the model will not call any tool. auto means the model can pick between generating a message or calling one or more tools. required means the model must call one or more tools. title: Tool Choice Mode - oneOf: - type: object properties: type: type: string enum: - function description: The type of tool choice. function: type: object properties: name: type: string description: The name of the function to require. required: - name description: Function tool choice details. required: - type - function additionalProperties: false title: Function - type: object properties: type: type: string enum: - custom description: The type of tool choice. custom: type: object properties: name: type: string description: The name of the custom tool to require. required: - name description: Custom tool choice details. required: - type - custom additionalProperties: false title: Custom - type: object properties: type: type: string enum: - allowed_tools allowed_tools: type: object properties: mode: type: string enum: - auto - required tools: type: array items: oneOf: - type: object properties: type: type: string enum: - function description: The type of the tool. function: type: object properties: name: description: The name of the function. type: string maxLength: 256 description: description: A description of what the function does. nullable: true type: string parameters: description: The parameters the function accepts. nullable: true type: object additionalProperties: {} strict: description: >- Whether to enable strict schema adherence when generating function calls. nullable: true default: false type: boolean required: - name description: The function definition. required: - type - function additionalProperties: false title: Function Tool - type: object properties: type: type: string enum: - custom description: The type of the tool. custom: type: object properties: name: type: string description: The name of the custom tool. description: description: >- A description of what the custom tool does. nullable: true type: string format: description: >- The format specification for the custom tool. nullable: true oneOf: - type: object properties: type: type: string enum: - text required: - type additionalProperties: false title: Text Format - type: object properties: type: type: string enum: - grammar grammar: type: object properties: syntax: type: string enum: - lark - regex definition: type: string required: - syntax - definition required: - type - grammar additionalProperties: false title: Grammar Format required: - name description: The custom tool definition. required: - type - custom additionalProperties: false title: Custom Tool required: - mode - tools required: - type - allowed_tools title: Allowed Tools tools: description: A list of tools the model may call. nullable: true type: array items: oneOf: - type: object properties: type: type: string enum: - function description: The type of the tool. function: type: object properties: name: description: The name of the function. type: string maxLength: 256 description: description: A description of what the function does. nullable: true type: string parameters: description: The parameters the function accepts. nullable: true type: object additionalProperties: {} strict: description: >- Whether to enable strict schema adherence when generating function calls. nullable: true default: false type: boolean required: - name description: The function definition. required: - type - function additionalProperties: false title: Function Tool - type: object properties: type: type: string enum: - custom description: The type of the tool. custom: type: object properties: name: type: string description: The name of the custom tool. description: description: A description of what the custom tool does. nullable: true type: string format: description: The format specification for the custom tool. nullable: true oneOf: - type: object properties: type: type: string enum: - text required: - type additionalProperties: false title: Text Format - type: object properties: type: type: string enum: - grammar grammar: type: object properties: syntax: type: string enum: - lark - regex definition: type: string required: - syntax - definition required: - type - grammar additionalProperties: false title: Grammar Format required: - name description: The custom tool definition. required: - type - custom additionalProperties: false title: Custom Tool top_logprobs: nullable: true type: integer minimum: 0 maximum: 20 description: >- An integer between 0 and 20 specifying the number of most likely tokens to return at each token position, each with an associated log probability. top_p: description: >- An alternative to sampling with temperature, called nucleus sampling, where the model considers the results of the tokens with top_p probability mass. nullable: true type: number minimum: 0 maximum: 1 user: description: >- A unique identifier representing your end-user, which can help OpenAI to monitor and detect abuse. nullable: true type: string verbosity: nullable: true type: string enum: - low - medium - high description: >- Constrains the verbosity of the model's response. Lower values will result in more concise responses, while higher values will result in more verbose responses. Currently supported values are low, medium, and high. web_search_options: description: Options for web search integration. nullable: true type: object properties: search_context_size: description: >- The amount of context to retrieve from web search results. nullable: true type: string enum: - low - medium - high user_location: nullable: true type: object properties: type: type: string enum: - approximate description: The location type. approximate: type: object properties: timezone: type: string nullable: true description: >- IANA timezone identifier (e.g. America/New_York, Europe/London, Asia/Tokyo). country: type: string nullable: true description: >- ISO 3166-1 alpha-2 country code (e.g. US, GB, DE, JP). city: type: string description: >- Free text input for the city of the user, e.g. San Francisco. nullable: true region: type: string description: >- Free text input for the region of the user, e.g. California. nullable: true description: Approximate location details. required: - type - approximate description: User location for web search. routing: nullable: true type: object properties: model: nullable: true type: object properties: fallbacks: nullable: true type: array items: description: One of 4631 allowed values. type: string description: >- Ordered list of models to fall back to if the primary model is unavailable or fails. Using auto in the fallbacks will move it to the end of the fallback list, and can be used to continuously try models until one succeeds. sort: nullable: true default: performance description: >- The metric used to sort and select among routing candidates, such as cost, performance, or a percentile latency metric (e.g. p95_latency). Defaults to performance. All metrics are cached for the most recent 15 minutes (except percentile latencies other than p50, p90, and p99). anyOf: - type: string enum: - cost - performance - total_requests - uptime - throughput - avg_latency - latency - avg_e2e_latency - e2e_latency - input_tokens - output_tokens - total_tokens - type: string pattern: ^p([0-9]|[1-9][0-9]|100)_(latency|e2e_latency)$ description: Model-level routing configuration. provider: nullable: true type: object properties: fallbacks: nullable: true type: array items: type: string description: One of 65 allowed values. description: >- Ordered list of providers to fall back to if the primary provider is unavailable or fails. sort: nullable: true default: performance description: >- The metric used to sort and select among routing candidates, such as cost, performance, or a percentile latency metric (e.g. p95_latency). Defaults to performance. All metrics are cached for the most recent 15 minutes (except percentile latencies other than p50, p90, and p99). anyOf: - type: string enum: - cost - performance - total_requests - uptime - throughput - avg_latency - latency - avg_e2e_latency - e2e_latency - input_tokens - output_tokens - total_tokens - type: string pattern: ^p([0-9]|[1-9][0-9]|100)_(latency|e2e_latency)$ interval: nullable: true description: >- The time window over which routing metrics are evaluated. Can be written as '1h' or '1 hour'. Defaults to 15 minutes. All metrics are cached for the most recent 15 minutes (except percentile latencies other than p50, p90, and p99). Any historical metrics longer than 15 minutes are then cached for 15 minutes upon request, and are refreshed when you use those metrics again. type: string pattern: >- ^\d+(?:m| minutes|h| hours?|d| days?|w| weeks?|y| years?)$ description: Provider-level routing configuration. description: >- Concentrate routing configuration controlling how requests are routed across models and providers. Learn more about routing. required: - messages - model responses: '200': description: Default Response content: application/json: schema: type: object properties: id: type: string model: type: string description: The model used for the chat completion. created: type: number minimum: 0 description: >- The Unix timestamp (in seconds) of when the chat completion was created. object: type: string description: >- The object type, which is always 'chat.completion' or 'chat.completion.chunk'. choices: type: array items: type: object properties: index: type: number minimum: 0 description: The index of the choice in the list of choices. logprobs: nullable: true type: object properties: content: default: [] description: >- Log probability information for the tokens in the response. type: array items: type: object properties: token: type: string bytes: type: array items: type: number logprob: type: number top_logprobs: type: array items: type: object properties: token: type: string bytes: type: array items: type: number logprob: type: number required: - token - bytes - logprob additionalProperties: false required: - token - bytes - logprob - top_logprobs additionalProperties: false refusal: default: [] description: >- Log probability information for tokens that were filtered or refusal tokens. type: array items: type: object properties: token: type: string bytes: type: array items: type: number logprob: type: number top_logprobs: type: array items: type: object properties: token: type: string bytes: type: array items: type: number logprob: type: number required: - token - bytes - logprob additionalProperties: false required: - token - bytes - logprob - top_logprobs additionalProperties: false required: - content - refusal additionalProperties: false description: Log probability information, if requested. finish_reason: type: string enum: - content_filter - function_call - length - stop - tool_calls description: The reason the model stopped generating tokens. message: type: object properties: role: type: string enum: - assistant description: The role of the author of this message. content: nullable: true description: The contents of the message. type: string reasoning: description: The reasoning content generated by the model. nullable: true type: string reasoning_content: description: >- The reasoning content generated by the model. Included for compatibility. `reasoning` will be returned rather than `reasoning_content`. nullable: true type: string tool_calls: description: Tool calls generated by the model, if any. nullable: true type: array items: oneOf: - type: object properties: type: type: string enum: - function description: The type of the tool call. id: type: string description: The ID of the tool call. function: type: object properties: name: description: The name of the function to call. arguments: type: string description: The arguments to pass to the function. required: - name - arguments additionalProperties: false description: The function that the model called. required: - type - id - function additionalProperties: false title: Function Tool Call - type: object properties: type: type: string enum: - custom description: The type of the tool call. id: type: string description: The ID of the tool call. custom: type: object properties: input: type: string description: The input to pass to the custom tool. name: description: The name of the custom tool. required: - input - name additionalProperties: false description: The custom tool that the model called. required: - type - id - custom additionalProperties: false title: Custom Tool Call annotations: description: Annotations for the message content. nullable: true type: array items: type: object properties: type: type: string enum: - url_citation description: The type of annotation. url_citation: type: object properties: url: type: string description: The URL being cited. title: type: string description: The title of the cited content. start_index: type: number minimum: 0 description: >- The start index of the citation in the content. end_index: type: number minimum: 0 description: >- The end index of the citation in the content. required: - url - title - start_index - end_index additionalProperties: false description: The URL citation details. required: - type - url_citation additionalProperties: false required: - role - content additionalProperties: false description: A chat completion message generated by the model. required: - index - finish_reason - message additionalProperties: false description: >- A list of chat completion choices. Can contain more than one if n is greater than 1. metadata: description: >- Set of 16 key-value pairs that can be attached to an object. This can be useful for storing additional information about the object in a structured format. nullable: true type: object additionalProperties: type: string maxLength: 512 service_tier: nullable: true type: string enum: - auto - default - flex - scale - priority description: >- Specifies the processing type used for serving the request. Determines the pricing and performance tier used to process the request. When not set, the default behavior is auto. Currently unsupported, but included for compatibility. usage: type: object properties: completion_tokens: description: Number of tokens in the generated completion. type: number minimum: 0 prompt_tokens: description: Number of tokens in the prompt. type: number minimum: 0 total_tokens: description: >- Total number of tokens used in the request (prompt + completion). type: number minimum: 0 prompt_tokens_details: type: object properties: audio_tokens: description: Audio tokens in the prompt. type: number minimum: 0 cached_tokens: type: number minimum: 0 description: Number of tokens in the prompt that were cached. cache_write_tokens: description: >- Number of tokens in the prompt that were written to cache. type: number minimum: 0 required: - cached_tokens additionalProperties: false description: Breakdown of prompt tokens. completion_tokens_details: nullable: true type: object properties: accepted_prediction_tokens: description: >- Number of tokens accepted from the prediction input. type: number minimum: 0 audio_tokens: description: Audio tokens in the completion. type: number minimum: 0 reasoning_tokens: type: number minimum: 0 description: Number of tokens used for reasoning. rejected_prediction_tokens: description: >- Number of tokens from the prediction input that were rejected. type: number minimum: 0 required: - reasoning_tokens additionalProperties: false description: Breakdown of completion tokens. additionalProperties: false description: Usage statistics for the completion request. cost: type: object properties: total: type: number byok: type: boolean description: >- Whether the request was served using one of your BYOK provider keys. breakdown: type: object additionalProperties: type: object properties: input_tokens: type: number minimum: 0 input_tokens_details: type: object properties: cached_tokens: type: number minimum: 0 description: >- The number of input tokens read from the prompt cache. cached_tokens_created: nullable: true type: number minimum: 0 description: >- Deprecated: use cache_write_tokens instead. The number of input tokens written to the prompt cache. Still populated for now, but will be removed in a future release. deprecated: true cache_write_tokens: nullable: true type: number minimum: 0 description: >- The number of input tokens written to the prompt cache. required: - cached_tokens additionalProperties: false output_tokens: type: number minimum: 0 output_tokens_details: type: object properties: reasoning_tokens: type: number minimum: 0 required: - reasoning_tokens additionalProperties: false tool_calls: nullable: true type: object properties: web_search: nullable: true type: number minimum: 0 exa_web_search: nullable: true type: number minimum: 0 additionalProperties: false total_tokens: type: number minimum: 0 required: - input_tokens - input_tokens_details - output_tokens - output_tokens_details - total_tokens additionalProperties: false required: - total - byok - breakdown additionalProperties: false description: The cost of the completion request. redact: nullable: true type: object properties: entities_found: type: number entity_types: type: array items: type: string models_used: type: array items: type: string execution_time_ms: type: number redaction_coverage: type: number redacted_input: type: array items: $ref: '#/components/schemas/InputItem' redacted_tools: type: array items: $ref: '#/components/schemas/Tool' additionalProperties: false required: - id - model - created - object - choices - usage - cost additionalProperties: false default: description: Default Response content: application/json: schema: anyOf: - type: object properties: error: type: object properties: code: type: string description: One of 25 allowed values. message: type: string upstream_error: type: string description: >- The raw JSON error payload returned by the upstream provider, present when the error originated there required: - code - message additionalProperties: false required: - error additionalProperties: false components: schemas: PromptCacheBreakpointInput: type: object properties: mode: type: string description: Creates an explicit prompt-cache breakpoint after this block. enum: - explicit required: - mode description: OpenAI explicit prompt-cache breakpoint settings. PromptCacheOptionsInput: type: object properties: mode: type: string enum: - implicit - explicit description: >- The prompt-caching mode. Explicit mode writes only at declared breakpoints. ttl: type: string enum: - 5m - 30m - 1h description: >- The minimum lifetime of cached prefixes written by this request. OpenAI supports 30m natively; 5m and 1h are accepted for cross-provider compatibility and snap to the nearest TTL the routed provider supports. Providers may retain cached prefixes longer. description: OpenAI prompt-cache behavior for this request. TextFormatInput: type: object properties: type: type: string description: The type of the format. Always text. enum: - text required: - type title: Text Format description: Default response format. Used to generate text responses. InputItem: anyOf: - $ref: '#/components/schemas/InputMessage' - $ref: '#/components/schemas/Item' description: 'A single input item: a message or an item.' Tool: oneOf: - $ref: '#/components/schemas/FunctionTool' - $ref: '#/components/schemas/WebSearchTool' - $ref: '#/components/schemas/CustomTool' description: A tool the model may call while generating a response. InputMessage: type: object properties: type: default: message description: The type of the input item. Always message. type: string enum: - message role: type: string enum: - user - assistant - system - developer description: The role of the message author. content: anyOf: - type: string minLength: 1 - type: array items: $ref: '#/components/schemas/InputItemContent' description: 'The content of the message: a string, or an array of content parts.' phase: type: string enum: - commentary - final_answer nullable: true required: - role - content additionalProperties: false title: Message description: A message input to the model. Item: anyOf: - $ref: '#/components/schemas/InputMessageItem' - $ref: '#/components/schemas/OutputMessageItem' - $ref: '#/components/schemas/FunctionToolCallItem' - $ref: '#/components/schemas/FunctionToolCallOutputItem' - $ref: '#/components/schemas/CustomToolCallItem' - $ref: '#/components/schemas/CustomToolCallOutputItem' - $ref: '#/components/schemas/ReasoningItem' - $ref: '#/components/schemas/WebSearchToolCallItem' - $ref: '#/components/schemas/CompactionItem' description: A single item in the model's input or output. FunctionTool: type: object properties: type: type: string description: The type of the tool. Always function. enum: - function name: {} description: type: string description: >- A description of the function. Used by the model to determine whether or not to call the function. nullable: true parameters: type: object additionalProperties: {} description: A JSON schema object describing the parameters of the function. strict: type: boolean description: Whether to enforce strict parameter validation. nullable: true cache_control: $ref: '#/components/schemas/CacheControl' required: - type - name - parameters additionalProperties: false title: Function Tool description: Defines a function in your own code the model can choose to call. WebSearchTool: type: object properties: type: type: string enum: - web_search - web_search_2025_08_26 - web_search_preview - web_search_preview_2025_03_11 description: The type of the web search tool. engine: oneOf: - type: object properties: type: type: string enum: - auto - native required: - type additionalProperties: false title: Auto / Native - type: object properties: type: type: string enum: - exa required: - type additionalProperties: false title: Exa nullable: true filters: type: object properties: allowed_domains: minItems: 1 maxItems: 1200 type: array items: type: string description: >- Allowed domains for the search. If not provided, all domains are allowed. nullable: true blocked_domains: minItems: 1 maxItems: 1200 type: array items: type: string description: >- Blocked domains for the search. If not provided, all domains are allowed. nullable: true additionalProperties: false description: Filters applied to the web search. nullable: true search_context_size: type: string enum: - low - medium - high description: >- High level guidance for the amount of context window space to use for the search. One of low, medium, or high. nullable: true user_location: allOf: - $ref: '#/components/schemas/UserLocation' nullable: true cache_control: $ref: '#/components/schemas/CacheControl' required: - type additionalProperties: false title: Web Search Tool description: Search the web for the latest information before generating a response. CustomTool: type: object properties: type: type: string description: The type of the tool. Always custom. enum: - custom name: {} description: type: string description: >- Optional description of the custom tool, used to provide more context. nullable: true format: allOf: - $ref: '#/components/schemas/CustomToolFormat' nullable: true cache_control: $ref: '#/components/schemas/CacheControl' required: - type - name additionalProperties: false title: Custom Tool description: A custom tool that processes input using a specified format. InputItemContent: oneOf: - $ref: '#/components/schemas/InputText' - $ref: '#/components/schemas/InputImage' - $ref: '#/components/schemas/InputFile' description: 'A single piece of content in an input message: text, image, or file.' InputMessageItem: type: object properties: type: default: message description: The type of the item. Always message. type: string enum: - message role: type: string enum: - user - system - developer description: The role of the message author. status: type: string enum: - completed - in_progress - incomplete description: >- The status of the item. One of in_progress, completed, or incomplete. nullable: true content: type: array items: $ref: '#/components/schemas/InputItemContent' description: The content parts of the message. required: - role - content additionalProperties: false title: Input Message description: A message item in the model's input. OutputMessageItem: type: object properties: type: default: message description: The type of the item. Always message. type: string enum: - message role: type: string description: The role of the message author. Always assistant. enum: - assistant id: type: string description: The unique ID of the output message. nullable: true status: type: string enum: - completed - in_progress - incomplete description: >- The status of the item. One of in_progress, completed, or incomplete. nullable: true content: type: array items: $ref: '#/components/schemas/OutputItemContent' description: The content parts of the message. cache_control: $ref: '#/components/schemas/CacheControl' phase: type: string enum: - commentary - final_answer nullable: true required: - role - content additionalProperties: false title: Output Message description: A message item produced by the model. FunctionToolCallItem: type: object properties: type: type: string description: The type of the item. Always function_call. enum: - function_call call_id: type: string description: >- The unique ID of the function tool call, used to correlate it with its output. name: {} arguments: type: string description: A JSON string of the arguments to pass to the function. id: type: string description: The unique ID of the function tool call item. nullable: true status: type: string enum: - completed - in_progress - incomplete description: >- The status of the item. One of in_progress, completed, or incomplete. nullable: true cache_control: $ref: '#/components/schemas/CacheControl' required: - type - call_id - name - arguments additionalProperties: false title: Function Tool Call description: A tool call to a function generated by the model. FunctionToolCallOutputItem: type: object properties: type: type: string description: The type of the item. Always function_call_output. enum: - function_call_output call_id: type: string minLength: 1 description: The unique ID of the function tool call this output is for. output: anyOf: - type: string - type: array items: $ref: '#/components/schemas/InputItemContent' description: The output of the function tool call. id: type: string format: uuid pattern: >- ^([0-9a-fA-F]{8}-[0-9a-fA-F]{4}-4[0-9a-fA-F]{3}-[89abAB][0-9a-fA-F]{3}-[0-9a-fA-F]{12})$ description: The unique ID of the function tool call output item. nullable: true status: type: string enum: - completed - in_progress - incomplete description: >- The status of the item. One of in_progress, completed, or incomplete. nullable: true cache_control: $ref: '#/components/schemas/CacheControl' required: - type - call_id - output additionalProperties: false title: Function Tool Call Output description: The output of a function tool call, provided back to the model. CustomToolCallItem: type: object properties: type: type: string description: The type of the item. Always custom_tool_call. enum: - custom_tool_call call_id: type: string description: >- The unique ID of the custom tool call, used to correlate it with its output. name: {} input: type: string description: The input passed to the custom tool. id: type: string description: The unique ID of the custom tool call item. nullable: true required: - type - call_id - name - input additionalProperties: false title: Custom Tool Call description: A call to a custom tool generated by the model. CustomToolCallOutputItem: type: object properties: type: type: string description: The type of the item. Always custom_tool_call_output. enum: - custom_tool_call_output call_id: type: string description: The unique ID of the custom tool call this output is for. output: type: string description: The output of the custom tool call. id: type: string description: The unique ID of the custom tool call output item. nullable: true status: type: string enum: - completed - in_progress - incomplete description: >- The status of the item. One of in_progress, completed, or incomplete. nullable: true required: - type - call_id - output additionalProperties: false title: Custom Tool Call Output description: The output of a custom tool call, provided back to the model. ReasoningItem: type: object properties: type: type: string description: The type of the item. Always reasoning. enum: - reasoning id: type: string description: The unique ID of the reasoning item. nullable: true summary: type: array items: $ref: '#/components/schemas/SummaryText' description: A summary of the reasoning performed by the model. content: type: array items: $ref: '#/components/schemas/ReasoningText' description: The reasoning text produced by the model. nullable: true encrypted_content: type: string description: >- The encrypted reasoning content, returned when reasoning is not stored. nullable: true status: type: string enum: - completed - in_progress - incomplete description: >- The status of the item. One of in_progress, completed, or incomplete. nullable: true required: - type - summary additionalProperties: false title: Reasoning description: A reasoning item produced by the model. WebSearchToolCallItem: type: object properties: type: type: string description: The type of the item. Always web_search_call. enum: - web_search_call id: type: string description: The unique ID of the web search tool call. action: $ref: '#/components/schemas/WebSearchToolCallAction' status: anyOf: - type: string enum: - completed - in_progress - incomplete description: >- The status of the item. One of in_progress, completed, or incomplete. - type: string enum: - searching description: The status of the web search tool call. nullable: true cache_control: $ref: '#/components/schemas/CacheControl' required: - type - id additionalProperties: false title: Web Search Tool Call description: A web search call generated by the model. CompactionItem: type: object properties: type: type: string description: The type of the item. Always compaction. enum: - compaction id: type: string description: The unique ID of the compaction item. nullable: true encrypted_content: type: string maxLength: 10485760 description: The encrypted compacted conversation context. required: - type - encrypted_content additionalProperties: false title: Compaction description: A compaction item containing compacted conversation context. CacheControl: type: object properties: type: default: ephemeral description: The cache control type. Currently only ephemeral is supported. type: string enum: - ephemeral ttl: default: 5m description: >- The time-to-live for the cache entry. One of 5m or 1h. Defaults to 5m. type: string enum: - 5m - 1h required: - type additionalProperties: false description: Cache control settings for explicit prompt caching of this content. UserLocation: type: object properties: type: type: string description: The type of location approximation. Always approximate. enum: - approximate timezone: type: string nullable: true description: >- IANA timezone identifier (e.g. America/New_York, Europe/London, Asia/Tokyo). country: type: string nullable: true description: ISO 3166-1 alpha-2 country code (e.g. US, GB, DE, JP). city: type: string description: Free text input for the city of the user, e.g. San Francisco. nullable: true region: type: string description: Free text input for the region of the user, e.g. California. nullable: true required: - type additionalProperties: false description: The approximate location of the user, used to refine web search results. CustomToolFormat: oneOf: - $ref: '#/components/schemas/TextFormat' - $ref: '#/components/schemas/GrammarFormat' description: >- The input format for a custom tool. Either unconstrained text or a grammar. InputText: type: object properties: type: type: string description: The type of the input item. Always input_text. enum: - input_text text: type: string minLength: 1 description: The text input to the model. cache_control: $ref: '#/components/schemas/CacheControl' prompt_cache_breakpoint: $ref: '#/components/schemas/PromptCacheBreakpoint' required: - type - text additionalProperties: false title: Input Text description: A text input to the model. InputImage: type: object properties: type: type: string description: The type of the input item. Always input_image. enum: - input_image image_url: type: string pattern: >- ^(https?:\/\/.+|data:image\/(png|jpe?g|webp|gif);base64,[A-Za-z0-9+/]+={0,2})$ description: >- The URL of the image. Either an http(s) URL or a base64-encoded data URL (data:image/;base64,...). detail: default: auto description: >- The detail level of the image to send to the model. One of low, high, auto, or original. Models that don't support 'original' convert it to 'high'. type: string enum: - low - high - auto - original nullable: true cache_control: $ref: '#/components/schemas/CacheControl' prompt_cache_breakpoint: $ref: '#/components/schemas/PromptCacheBreakpoint' required: - type - image_url additionalProperties: false title: Input Image description: An image input to the model. InputFile: type: object properties: type: type: string description: The type of the input item. Always input_file. enum: - input_file file_data: type: string pattern: ^(data:application\/pdf;base64,) description: >- The content of the file as a base64-encoded PDF data URL (data:application/pdf;base64,...). file_url: type: string pattern: ^https?:\/\/ description: The URL of the file to send to the model. filename: type: string description: The name of the file, only provided with file_url. cache_control: $ref: '#/components/schemas/CacheControl' prompt_cache_breakpoint: $ref: '#/components/schemas/PromptCacheBreakpoint' required: - type additionalProperties: false title: Input File description: A file input (PDF) to the model. OutputItemContent: oneOf: - $ref: '#/components/schemas/OutputText' description: A single piece of content in an output message. SummaryText: type: object properties: type: type: string description: The type of the content. Always summary_text. enum: - summary_text text: type: string description: The summarized reasoning text. required: - type - text additionalProperties: false description: A chunk of summarized reasoning text produced by the model. ReasoningText: type: object properties: type: type: string description: The type of the content. Always reasoning_text. enum: - reasoning_text text: type: string description: The reasoning text produced by the model. required: - type - text additionalProperties: false description: A chunk of reasoning text produced by the model. WebSearchToolCallAction: oneOf: - $ref: '#/components/schemas/WebSearchToolCallSearchAction' - $ref: '#/components/schemas/WebSearchToolCallOpenPageAction' - $ref: '#/components/schemas/WebSearchToolCallFindAction' description: An action performed by the web search tool. TextFormat: type: object properties: type: type: string description: The type of the format. Always text. enum: - text required: - type additionalProperties: false title: Text Format description: Default response format. Used to generate text responses. GrammarFormat: type: object properties: type: type: string description: The type of the format. Always grammar. enum: - grammar syntax: type: string enum: - lark - regex description: The syntax of the grammar definition. One of lark or regex. definition: type: string description: The grammar definition. required: - type - syntax - definition additionalProperties: false title: Grammar Format description: A grammar format for a custom tool, defined using Lark or regex syntax. PromptCacheBreakpoint: type: object properties: mode: type: string description: Creates an explicit prompt-cache breakpoint after this block. enum: - explicit required: - mode additionalProperties: false description: OpenAI explicit prompt-cache breakpoint settings. OutputText: type: object properties: type: type: string description: The type of the output content. Always output_text. enum: - output_text text: type: string description: The text output from the model. annotations: type: array items: $ref: '#/components/schemas/Annotation' description: Annotations on the output text, such as URL citations. nullable: true logprobs: type: array items: type: object properties: token: type: string bytes: type: array items: type: number logprob: type: number top_logprobs: type: array items: type: object properties: token: type: string bytes: type: array items: type: number logprob: type: number required: - token - bytes - logprob additionalProperties: false required: - token - bytes - logprob - top_logprobs additionalProperties: false description: Log probability information for the output tokens. nullable: true cache_control: $ref: '#/components/schemas/CacheControl' prompt_cache_breakpoint: $ref: '#/components/schemas/PromptCacheBreakpoint' required: - type - text additionalProperties: false title: Output Text description: A text output from the model. WebSearchToolCallSearchAction: type: object properties: type: type: string description: The type of the action. Always search. enum: - search query: type: string description: The search query. queries: type: array items: type: string description: The list of search queries run. nullable: true sources: type: array items: $ref: '#/components/schemas/Source' description: The sources consulted during the search. nullable: true required: - type additionalProperties: false title: Search Action description: A search action performed by the web search tool. WebSearchToolCallOpenPageAction: type: object properties: type: type: string description: The type of the action. Always open_page. enum: - open_page url: type: string description: The URL of the page opened. nullable: true required: - type additionalProperties: false title: Open Page description: An action where the web search tool opens a page. WebSearchToolCallFindAction: type: object properties: type: type: string description: The type of the action. Always find_in_page. enum: - find_in_page pattern: type: string description: The pattern searched for within the page. url: type: string description: The URL of the page searched. required: - type - pattern - url additionalProperties: false title: Find In Page description: An action where the web search tool searches within a page. Annotation: oneOf: - $ref: '#/components/schemas/UrlCitation' description: An annotation on the output text. Source: oneOf: - $ref: '#/components/schemas/UrlSource' - $ref: '#/components/schemas/ApiSource' description: A source consulted by the web search tool. UrlCitation: type: object properties: type: type: string description: The type of the annotation. Always url_citation. enum: - url_citation start_index: type: integer minimum: 0 maximum: 9007199254740991 description: The index of the first character of the citation in the output text. end_index: type: integer minimum: 0 maximum: 9007199254740991 description: The index of the last character of the citation in the output text. title: type: string description: The title of the cited web page. url: type: string description: The URL of the cited web page. required: - type - start_index - end_index - title - url additionalProperties: false title: URL Citation description: A citation to a URL in the output text. UrlSource: type: object properties: type: type: string description: The type of the source. Always url. enum: - url url: type: string description: The URL of the source. required: - type - url additionalProperties: false title: URL Source description: A web URL source consulted by the web search tool. ApiSource: type: object properties: type: type: string description: The type of the source. Always api. enum: - api name: type: string description: The name of the API source. required: - type - name additionalProperties: false title: API Source description: An API source consulted by the web search tool. ````