A reference for Google's Developer Knowledge API and its official Model Context Protocol server, which provide a canonical, machine-readable gateway to Google's public developer documentation across Firebase, Android, and Google Cloud. Connecting the MCP server lets an AI assistant retrieve fresh documentation as Markdown, replacing outdated training data and fragile scraping with an authoritative source for implementation and troubleshooting.
Google has released a public preview of the Developer Knowledge API and an official MCP server built on it. Together they give AI agents and developer tools a canonical, machine-readable path to Google’s official documentation — so an assistant answers from current docs instead of stale training data.
The Developer Knowledge API
The Developer Knowledge API is meant to be the programmatic source of truth for Google’s public documentation. Rather than scraping pages or trusting whatever the model absorbed at training time, an agent searches for the right pages and retrieves them as Markdown.
- Coverage: documentation across Firebase, Android, Google Cloud, and more.
- Search and retrieve: find relevant pages, then pull their full Markdown content.
- Freshness: during the preview, pages are re-indexed within 24 hours of an update.
The official MCP server
The MCP server exposes that API to any compatible IDE or assistant, letting the agent “read” Google’s docs in real time. That supports workflows such as:
- Implementation guidance — “What’s the recommended way to implement push notifications in Firebase?”
- Troubleshooting — “Check the docs for how to fix the
ApiNotActivatedMapErrorin the Maps API.” - Comparison — “Compare Cloud Run and Cloud Functions for this use case.”
Getting started
- Create an API key. Generate and restrict a key for the Developer Knowledge API on your project’s Credentials page.
- Enable the MCP server with the Google Cloud CLI:
bash
gcloud beta services mcp enable developerknowledge.googleapis.com --project=PROJECT_ID - Configure your tool. Add the server to your assistant’s configuration file (for example,
mcp_config.json), following the official documentation.
Direction
The preview serves unstructured Markdown. Planned work includes structured content — code-sample objects and API-reference entities — along with a wider documentation corpus and lower re-indexing latency.
See also
- Developer Knowledge API
- Model Context Protocol (MCP)
- Google Cloud
- Firebase
- Android
- Agentic Tools
- Documentation API
Ask Lynx Assist
More Guides
Run disciplined SEO A/B tests in seven steps — one metric, two variations, randomized segments, run to significance, track, analyze the winner, and iterate.
Build a topic cluster in seven steps — select and score a pillar, validate it, map subtopics, align to intent, architect internal links, publish, and measure.
Prepare your site for AI search in five steps — content architecture, entity consistency, E-E-A-T, structured data, and machine-readable structure.
Get your content cited by AI in seven steps — answer capsules, link-free extraction, original data, digital PR, community presence, consistent messaging, and tracking.
A seven-step walkthrough for setting up Google Search Console on a new site — property type, DNS verification, sitemap, GA4 link, users, URL checks, and a monitoring routine.