API documentation
- OpenAPI 3.1 specification
Operations, parameters, response schemas, and structured errors.
- Search API
Search public Grail Computer pages and return ranked sources.
- Answer API
Return a concise answer with relevant public sources.
Developers and agents
Grail Computer developer resources
Use these public interfaces to discover, search, and cite Grail Computer content. The discovery API is read-only and does not require authentication.
Agent discovery
- llms.txt
Curated Markdown index for agents and language models.
- NLWeb discovery
JSON metadata, capabilities, endpoints, and suggested queries.
- Sitemap
Machine-readable inventory of public pages.
- Robots policy
Crawler access policy for search and AI agents.
- Agentic resource catalog
Typed index of Grail Computer agent resources and APIs.
- Agent Skills index
Conformant discovery index for the Grail public-discovery skill.
- Public-content MCP server card
Identity, transport, and tool metadata for the read-only MCP server.
- API llms.txt
Scoped agent index for stable REST and MCP operations.
- NLWeb Schema Map
Schema Feed map for the public JSON Lines knowledge feed.
Integration status
Grail Computer publishes a read-only public-content MCP server. It does not expose customer-system actions, private project data, or webhooks. The OpenAPI document is the source of truth for REST operations; versioned paths are stable, and deprecated aliases receive Deprecation and successor-version headers before removal.