Google Announces Developer Knowledge API and MCP Server for Direct Access to Official Docs from Generative AI
Contents
Google has announced a public preview of the Developer Knowledge API and its accompanying MCP (Model Context Protocol) Server. It is a practical addition for developers: generative AI tools can now look up Google’s official documentation directly instead of relying only on training data.
What it can do
With the Developer Knowledge API, you can programmatically retrieve official docs for Google technologies such as Google Cloud, Android, Firebase, Apigee, and web platform topics. Typical uses include searching related pages and fetching entire pages in Markdown format.
The MCP Server acts as a bridge between that API and AI assistants. Expected use cases include:
- Asking about Firebase push notifications and getting answers based on official guidance
- Troubleshooting errors such as Maps API’s
ApiNotActivatedMapErrorby referring to the official docs - Consulting an assistant about choices like whether Cloud Run or Cloud Functions is the better fit
Keeping up with documentation changes
Documentation updates are indexed within 24 hours. Because the assistant can consult the latest docs directly instead of depending only on model training data, the risk of surfacing stale information or deprecated APIs is lower.
Broader MCP adoption
MCP is the protocol proposed by Anthropic as a standard way for AI assistants to connect to external tools and data sources. Google’s decision to ship an official MCP Server is significant. It should make these docs directly usable from MCP-aware tools such as Claude Code and Cursor.
Once a documentation corpus as large as Google’s becomes accessible through MCP, AI-assisted development workflows get materially more reliable. It is still only a public preview, so production use may come later, but it already looks worth testing.
Source: ITmedia News