Overview
The Developer Knowledge API is Google official programmatic source of truth for developer documentation, launched on October 7, 2026 and covering Google Cloud, Firebase, Android and more. It replaces two things that have long caused friction: brittle web scraping for documentation, and the staleness that comes from LLM training cutoffs.
It takes the form of a structured API that returns Markdown-formatted documentation directly, on top of which it offers semantic and keyword search, intelligent document chunking and grounded question answering. The index is refreshed frequently to keep the lag between an upstream documentation change and what an agent can retrieve as short as possible.
There is more than one way in: a gcloud CLI surface, an official agent skill for AI coding assistants, an API Explorer for trying requests interactively, and multi-language client libraries. Whether you are typing in a terminal, letting an agent look things up itself, or calling from backend code, you reach the same official content.
Key Features
- Markdown documentation from the official source: A structured API serves Markdown-formatted documentation for Google Cloud, Firebase, Android and more, replacing scraping approaches that break whenever a page layout changes, and freeing agents from relying on stale versions baked into training data
- Semantic and keyword search together: Both semantic search and keyword search are supported, paired with intelligent document chunking, so you can find passages by meaning as well as hit exact interface or parameter names
- Grounded question answering: Grounded Q&A answers directly from official documentation content, so responses carry a traceable source and the mechanism itself suppresses fabrication from memory
- gcloud CLI works out of the box: The command-line surface comes pre-installed in Cloud Shell and works across standard Google Cloud SDK installations on Linux, macOS and Windows with no extra plugins or configuration. answer-query returns grounded answers, documents search-chunks searches snippets, and documents describe retrieves metadata and content for a specific document
- Standard MCP integration with major coding tools: The official agent skill connects over standard MCP to Google Antigravity, Claude Code, Cursor, GitHub Copilot and custom agentic frameworks, installs into a project with a single npx skills add command, and can fall back to the REST API via curl
- Multi-step retrieval saves tokens: The skill runs a multi-step retrieval workflow, inspecting code and documentation chunks with searchDocumentChunks before fetching complete Markdown pages with documents.get, or delivering grounded answers through answerQuery. That coarse-to-fine order keeps token and context use low
Use Cases
- [object Object]
- [object Object]
- [object Object]
- [object Object]
- [object Object]
Pros
- Documentation comes from the official source, so no scraping scripts to maintain and no breakage when page layouts change
- Markdown output is natively suited for feeding models and agents, removing an HTML cleaning step
- Grounded Q&A carries sources, which lowers hallucination by mechanism rather than by prompting
- The gcloud CLI ships pre-installed in Cloud Shell and with the standard SDK on all three desktop platforms, so onboarding is configuration-free
- Standard MCP coverage spans Antigravity, Claude Code, Cursor, Copilot and more
- A built-in chunks-first, full-page-second retrieval flow keeps token and context spend down
- Frequent index refreshes keep the delay between an upstream change and retrievability short
Pricing
Provided as Google official programmatic entry point for developer documentation. The companion gcloud CLI ships pre-installed in Cloud Shell and with the standard Google Cloud SDK, and the official agent skill installs free via npx skills add. Usage follows each product authentication model, supporting Application Default Credentials and API keys.
Summary
The Developer Knowledge API removes a very practical friction: when AI coding tools write Google Cloud or Firebase code, their basis is either an outdated version in training data or unstable web scraping. Google turned its own documentation into a structured API, delivered as Markdown, and shipped a CLI, an agent skill and multi-language client libraries alongside it, effectively making official documentation a first-class thing that programs can consume.
One design detail is worth noticing: the official skill deliberately requires searching chunks first and fetching full pages only when needed. That order keeps token spend predictable and shows the tooling was designed around real coding workflows rather than wrapping a documentation API and stopping there.
For anyone developing on the Google stack daily, the immediate gain is coding agents working from primary specifications. For teams building developer tools, it adds an official data source that can be embedded into a product.
Version History
- Developer Knowledge API ecosystem launches (2026-10-07): Google introduced this programmatic official source for developer documentation covering Google Cloud, Firebase, Android and more, delivered through a structured API in Markdown format to replace web scraping and sidestep staleness from model training cutoffs. It offers semantic and keyword search, intelligent document chunking and grounded question answering, with the index refreshed frequently. The ecosystem includes a gcloud CLI surface pre-installed in Cloud Shell with answer-query, documents search-chunks and documents describe, an official agent skill installed via npx skills add google/skills that reaches Antigravity, Claude Code, Cursor, GitHub Copilot and custom frameworks over standard MCP, plus an API Explorer and multi-language client libraries, using a chunks-first retrieval flow to save tokens