Endpoints
Messages
Create Anthropic Messages-compatible output with Capriole AI Claude models.
POST
Use this endpoint when your application, SDK, or agent expects the Anthropic Messages request and response format. Send a standard Messages payload withDocumentation Index
Fetch the complete documentation index at: https://docs.capriole.ai/llms.txt
Use this file to discover all available pages before exploring further.
model, messages, and max_tokens. Capriole AI rewrites the model to the selected upstream Claude model, forwards compatible fields such as system, tools, thinking, and stream, and returns the upstream JSON or SSE stream unchanged.
POST /v1/messages/count_tokens is also available for Anthropic-compatible clients that need a token-count preflight. It accepts the same message shape and returns the upstream token-count JSON unchanged.
For simple Capriole-native text generation, use POST /v1/chat. Use POST /v1/messages when Anthropic Messages wire behavior is required.Authorizations
Use an API key created in the Capriole AI page. Send it as Authorization: Bearer sk-....
Body
application/json
Public Anthropic Messages model identifier
Available options:
claude-opus-4-8, claude-opus-4-7 Anthropic-compatible conversation messages.
Maximum number of output tokens.
Optional Anthropic-compatible system prompt.
Anthropic-compatible tool definitions.
Anthropic-compatible thinking options.
Stream the response as Anthropic-compatible server-sent events.
Response
Anthropic Messages-compatible JSON response or SSE stream
Anthropic Messages-compatible response body.