Skip to main content
Sail provides OpenAI- and Anthropic-compatible inference APIs. You can use the OpenAI SDK, Anthropic SDK, or raw HTTP clients.
Alpha — Chat Completions and Messages API support is still in active development. Behavior and supported fields may change without notice. We recommend the Responses API for production workloads.

​
Authentication

Include your API key as a Bearer token in the Authorization header of every request: 
Authorization: Bearer YOUR_SAIL_API_KEY

​
Supported endpoints

APIEndpointStatus
ModelsGET /v1/modelsSupported
OpenAI ResponsesPOST /v1/responsesSupported
OpenAI ResponsesGET /v1/responses/{response_id}Supported
OpenAI Chat CompletionsPOST /v1/chat/completionsSupported
Anthropic MessagesPOST /v1/messagesSupported
Not currently supported:
  • DELETE /v1/responses/{response_id}
  • POST /v1/responses/{response_id}/cancel
  • POST /v1/messages/batches and related batch endpoints

​
Shared request metadata

All three inference APIs support metadata with string values. These keys are especially useful:
  • completion_window: "asap", "15m", or "24h"
  • completion_webhook: webhook URL for completion notification
  • webhook_token: optional bearer token that Sail includes when calling your webhook

​
Responses API

Supported endpoints:
  • POST /v1/responses
  • GET /v1/responses/{response_id}
Supported fields include:
  • model, input
  • max_output_tokens, temperature, top_p
  • text.format with:
    • {"type":"text"}
    • {"type":"json_schema", "name": "...", "schema": {...}}
  • reasoning.effort (low, medium, high)
  • reasoning.generate_summary (auto, concise, detailed)
  • background
  • metadata
  • user
Current limitations:
  • Text-only input (no image/audio/file input blocks)
  • stream=true is not supported
  • Tools are not supported (tools, parallel_tool_calls, custom tool_choice)
  • store=false is not supported
  • truncation only supports "disabled"

​
Chat Completions API

Supported endpoint:
  • POST /v1/chat/completions
Supported fields include:
  • model, messages
  • temperature, top_p
  • max_completion_tokens
  • response_format with text or json_schema
  • reasoning_effort
  • metadata
  • user
Current limitations:
  • Text-only message content (no image/audio content parts)
  • stream=true is not supported
  • Tools/function-calling fields are not supported
  • n must be 1 when provided

​
Messages API

Supported endpoint:
  • POST /v1/messages
Supported fields include:
  • model, max_tokens, messages
  • temperature, top_p
  • output_config.format with type: "json_schema"
  • metadata
Current limitations:
  • Text-only content blocks
  • stream=true is not supported
  • system, tools, tool_choice, top_k, thinking, and service_tier are not supported
Explore the full endpoint-level request and response schemas in the API reference tab.