Live docs
The live docs describe every feature of Agent from its code: the screens, the operations behind them, the API reference and the MCP tools. Each release builds them again.
What the docs showLink to What the docs show
Live docs has one page for each feature. A feature page shows:
- what the feature does, from its help article;
- the operations its screens call, and who may call each one;
- its screens, the events it writes to the audit log and the related features.
The API reference lists every operation of the API. The MCP tools page lists the tools that Claude, ChatGPT or Codex can call. The pages and operations for platform operators are not in the public docs.
Where the docs come fromLink to Where the docs come from
Each release reads the code and writes the docs again. Each page names the commit it came from. A page is never older than the release that serves it.
SearchLink to Search
Type in the search field on Live docs. The search looks at titles, summaries, screens and operation names.
The code viewLink to The code view
The platform team can open Code view for the platform team on a feature page. The code view shows the files, symbols, operations, tables and tests of the feature. Each one has its line in the code.
The platform team is the organization that runs this server. Its members, with any role, can open the code view. People in other organizations see a message instead.
For agents and toolsLink to For agents and tools
The same docs are JSON files under /docs-data/. index.json lists the features, and api.json and mcp.json hold the API and the MCP tools.
Checked against the product on 2026-10-14.
Was this helpful?
Related articles
- Use Agent from Claude, ChatGPT or CodexAdd Agent to Claude, ChatGPT, Claude Code or Codex. The model runs there on your plan and uses Agent's tools to read and assign work.
- API tokensAPI tokens are machine credentials for CI, scripts and MCP clients. Each token has one role and expires.
- IntegrationsIntegrations connect workers to other tools. Each integration has a status and a risk label, and you choose which workers may use it.