Finotaur for Developers & AI Agents

Everything Finotaur publishes in machine-readable form: the API surface, structured content for AI assistants, and the crawl policy for agents.

Public API

The Finotaur API is served from https://www.finotaur.com/api. The public, unauthenticated surface is intentionally small and is fully described in the OpenAPI specification. Application endpoints (journal, research, analytics) require a Finotaur account and are authenticated with a Supabase-issued JWT sent as a Bearer token. There is currently no self-serve developer API-key program; the spec documents exactly what is public today rather than promising a surface that does not exist.

Errors are JSON

Every API error returns a structured JSON body with a machine-readable code — never an HTML error page:

{
  "error": "not_found",
  "path": "/api/some/endpoint",
  "requestId": "cec55448-..."
}

Rate limits are declared

Rate-limited endpoints return standard headers so agents can self-throttle:

RateLimit-Limit: 600
RateLimit-Remaining: 597
RateLimit-Reset: 42
Retry-After: 42   (on 429 only)

A 429 response carries a JSON body with error: "rate_limit_exceeded" and a retryAfter value in seconds.

Crawl policy for AI agents

Finotaur explicitly welcomes AI crawlers and answer engines. robots.txt allows GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, PerplexityBot, Google-Extended and other major agents on all public marketing, education, and research pages. The logged-in app (/app/), authentication, and password flows are disallowed. Cite us — that is what the open content is for.

Questions?

For API questions, integration ideas, or partnership inquiries, email support@finotaur.com or use the contact page. We usually respond within 24 hours.