> ## Documentation Index > Fetch the complete documentation index at: https://docs.gloo.com/llms.txt > Use this file to discover all available pages before exploring further. # Anthropic Messages API Reference > Send Anthropic Messages API requests to Claude models with the official Anthropic SDKs, without guardrails. > **[Guide: Anthropic Messages](/api-guides/anthropic-messages)** > > SDK setup, streaming, thinking, tools, supported betas, limits and errors. ## OpenAPI ````yaml post /ai/v2/direct/anthropic/v1/messages openapi: 3.1.0 info: title: AI API description: |- Optimized for high-speed inference and scalability by Gloo AI [Swagger UI](./docs) | [ReDoc UI](./redoc) version: 1.0.0 servers: - url: https://platform.ai.gloo.com security: - bearerAuth: [] tags: - name: Content Controls - name: Data Engine - name: Ingestion & Enrichment - name: answers - name: chat - name: completions - name: content - name: core - name: data - name: get-models - name: ingestion - name: sotc paths: /ai/v2/direct/anthropic/v1/messages: post: tags: - v2 summary: Anthropic Messages description: >- Anthropic's Messages API for Claude models, for use with the official Anthropic SDKs. Set the SDK's base URL to `https://platform.ai.gloo.com/ai/v2/direct/anthropic` and pass your Gloo API key as the auth token (`Authorization: Bearer`). The request body and the response, streaming events and errors follow Anthropic's Messages API. Claude models only; no guardrails, output moderation or routing are applied. operationId: post-v2-direct-anthropic-messages parameters: - name: anthropic-version in: header required: false description: Optional. If sent, must be `2023-06-01`. schema: type: string enum: - '2023-06-01' - name: anthropic-beta in: header required: false description: >- Comma-separated beta features. Supported: `context-management-2025-06-27`, `thinking-display-updates-2026-08-18`, `interleaved-thinking-2025-05-14`, `fine-grained-tool-streaming-2025-05-14`. Any other value is dropped and named in `x-gloo-ignored-betas`. schema: type: string requestBody: required: true content: application/json: schema: $ref: '#/components/schemas/AnthropicMessagesRequest' responses: '200': description: >- The model's message. With `stream: true`, a Server-Sent Events stream of Anthropic's Messages events (`message_start`, `content_block_start`, `content_block_delta`, `content_block_stop`, `message_delta`, `message_stop`, `ping`). A failure after the stream starts ends it with an `error` event. headers: request-id: description: >- The request's ID, in the form `req_...`. Include it when you contact support. schema: type: string x-gloo-ignored-params: description: >- Request fields that were dropped, comma-separated: top-level body fields, and `messages.thinking` when earlier `thinking` or `redacted_thinking` blocks were removed from `messages`. schema: type: string x-gloo-ignored-betas: description: '`anthropic-beta` values that were dropped, comma-separated.' schema: type: string x-gloo-adjusted-params: description: >- Values Gloo changed, such as `max_tokens=128000` when `max_tokens` was above the model's output limit. schema: type: string content: application/json: schema: $ref: '#/components/schemas/AnthropicMessage' text/event-stream: schema: type: string description: >- Anthropic Messages streaming events, passed through as Claude sends them. '400': description: >- `invalid_request_error`: the request is invalid, uses an unsupported tool, content block, source or `anthropic-version`, asks for `max_tokens` above 21,333 without streaming, or contains a phrase blocked on this platform. headers: request-id: description: >- The request's ID, in the form `req_...`. Include it when you contact support. schema: type: string content: application/json: schema: $ref: '#/components/schemas/AnthropicMessagesError' example: type: error error: type: invalid_request_error message: >- tools.0.type: "web_fetch_20250910" is not supported on this endpoint. details: code: tool_not_supported retryable: false request_id: req_... '401': description: The API key is missing or invalid. '402': description: '`billing_error`: the organization has no credit left.' headers: request-id: description: >- The request's ID, in the form `req_...`. Include it when you contact support. schema: type: string content: application/json: schema: $ref: '#/components/schemas/AnthropicMessagesError' '403': description: >- `permission_error`: the API key's organization can't make this request. headers: request-id: description: >- The request's ID, in the form `req_...`. Include it when you contact support. schema: type: string content: application/json: schema: $ref: '#/components/schemas/AnthropicMessagesError' '404': description: >- `not_found_error`: the model is not a Claude model (`model_not_available`), the model is retired (`model_retired`, the message names its replacement), or the path is not served (`endpoint_not_available`). headers: request-id: description: >- The request's ID, in the form `req_...`. Include it when you contact support. schema: type: string content: application/json: schema: $ref: '#/components/schemas/AnthropicMessagesError' '413': description: '`request_too_large`: the request body is over 32 MB.' headers: request-id: description: >- The request's ID, in the form `req_...`. Include it when you contact support. schema: type: string content: application/json: schema: $ref: '#/components/schemas/AnthropicMessagesError' '429': description: >- `rate_limit_error`: too many requests in flight for the organization (`retry-after: 1`), or the organization's spending limit is reached (`spend_limit_reached`, with `x-should-retry: false`). headers: request-id: description: >- The request's ID, in the form `req_...`. Include it when you contact support. schema: type: string retry-after: description: Seconds to wait before retrying. schema: type: integer x-should-retry: description: >- `false` when retrying the same request will not help. The Anthropic SDKs follow it. schema: type: string content: application/json: schema: $ref: '#/components/schemas/AnthropicMessagesError' '500': description: >- `api_error`: Gloo could not complete the request. Retry with backoff. headers: request-id: description: >- The request's ID, in the form `req_...`. Include it when you contact support. schema: type: string content: application/json: schema: $ref: '#/components/schemas/AnthropicMessagesError' '504': description: >- `timeout_error`: a non-streaming request ran too long. Use streaming. headers: request-id: description: >- The request's ID, in the form `req_...`. Include it when you contact support. schema: type: string x-should-retry: description: >- `false` when retrying the same request will not help. The Anthropic SDKs follow it. schema: type: string content: application/json: schema: $ref: '#/components/schemas/AnthropicMessagesError' '529': description: >- `overloaded_error`: the model is temporarily unavailable, or the stream did not start in time. Retry after `retry-after`. headers: request-id: description: >- The request's ID, in the form `req_...`. Include it when you contact support. schema: type: string retry-after: description: Seconds to wait before retrying. schema: type: integer content: application/json: schema: $ref: '#/components/schemas/AnthropicMessagesError' example: type: error error: type: overloaded_error message: The model is temporarily unavailable. Retry shortly. details: code: upstream_capacity retryable: true request_id: req_... components: schemas: AnthropicMessagesRequest: type: object title: AnthropicMessagesRequest description: >- An Anthropic Messages request. Fields are as defined by Anthropic's Messages API. Any other top-level field is dropped and named in `x-gloo-ignored-params`. required: - model - max_tokens - messages properties: model: type: string description: >- A Claude model: a Gloo model ID such as `gloo-anthropic-claude-opus-5.5`, or Anthropic's ID such as `claude-opus-5-5`. example: gloo-anthropic-claude-opus-5.5 max_tokens: type: integer minimum: 1 description: >- Maximum output tokens, including thinking. Above the model's output limit it is lowered to that limit. Above 21,333 requires `stream: true`. messages: type: array minItems: 1 description: >- The conversation. Roles are `user` and `assistant`. Content is a string or an array of blocks: `text`, `image` (base64 or URL source), `document` (base64, text, content or URL source), `search_result`, `tool_use`, `tool_result`, `thinking`, `redacted_thinking`, and `server_tool_use` / `web_search_tool_result` when the web search tool is in `tools`. `thinking` and `redacted_thinking` blocks are accepted but not passed to the model; the `x-gloo-ignored-params` header lists `messages.thinking` when any were sent. items: type: object required: - role - content properties: role: type: string enum: - user - assistant content: anyOf: - type: string - type: array items: type: object system: description: 'System prompt: a string or an array of text blocks.' anyOf: - type: string - type: array items: type: object stream: type: boolean description: Stream the response as Server-Sent Events. default: false stop_sequences: type: array items: type: string temperature: type: number top_p: type: number top_k: type: integer tools: type: array description: >- Your own tools, plus the Anthropic-defined web search (`web_search_*`), bash (`bash_*`) and text editor (`text_editor_*`) tools. Other tool types return `400`. items: type: object tool_choice: type: object thinking: type: object description: >- Thinking configuration, sent as written, for example `{"type": "adaptive", "display": "summarized"}`. output_config: type: object description: >- Output configuration, sent as written, for example `{"effort": "high"}`. cache_control: type: object description: 'Automatic prompt caching, for example `{"type": "ephemeral"}`.' context_management: type: object description: >- Context editing. Requires the `context-management-2025-06-27` beta; otherwise dropped. metadata: type: object description: >- Only `user_id` is used, and it is not passed to the model provider as sent. properties: user_id: type: string AnthropicMessage: type: object title: AnthropicMessage description: An Anthropic `message` object. `model` is the model ID you sent. properties: id: type: string type: type: string enum: - message role: type: string enum: - assistant model: type: string content: type: array items: type: object stop_reason: type: - string - 'null' stop_sequence: type: - string - 'null' usage: type: object properties: input_tokens: type: integer output_tokens: type: integer cache_creation_input_tokens: type: - integer - 'null' cache_read_input_tokens: type: - integer - 'null' AnthropicMessagesError: type: object title: AnthropicMessagesError description: >- Anthropic's error shape. Show `error.message` to users; branch on `error.details.code`. required: - type - error properties: type: type: string enum: - error error: type: object required: - type - message properties: type: type: string enum: - invalid_request_error - authentication_error - billing_error - permission_error - not_found_error - request_too_large - rate_limit_error - api_error - timeout_error - overloaded_error message: type: string description: A short message safe to show your users. details: type: object properties: code: type: string description: A stable code to branch on, such as `model_retired`. developer_message: type: string description: What to change, when there is more to say. retryable: type: boolean request_id: type: string securitySchemes: bearerAuth: type: http scheme: bearer bearerFormat: api-key x-default: description: >- Bearer authentication header of the form Bearer ``, where `` is your [API key](/studio/manage-api-credentials). ```` This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.