Mirroring the Gemini API documentation in clean, structured Markdown.
An offline-first, LLM-optimized collection of the Gemini API documentation. Every capability, every model, and every guide — organized and ready for your next project.
Note: read structure.md to understand the full hierarchy and available categories.
- 🏎️ Optimized for LLMs: Perfectly formatted for RAG (Retrieval Augmented Generation) and AI context.
- 📂 Structured Layout: Mirrors the official Gemini API Documentation.
- 💻 Multi-Language: Comprehensive code snippets for
Python,JavaScript,Go,Java,C#,REST/curl, andApps Script. - 🔍 Searchable: Grep your way through the entire Gemini ecosystem in milliseconds.
- 📄 Clean Markdown: Zero fluff, just the docs.
The documentation is organized across two main directories:
docs/
├── get_started/ # Quickstart, API keys, Libraries, Pricing...
├── core_capabilities/ # Text, Image, Video, Audio, Function Calling, Thinking...
├── models/ # Gemini 3, Embeddings, Imagen, Veo, Lyria, TTS...
├── tools_and_agents/ # Code Execution, Computer Use, Deep Research, Search...
├── live_api/ # Real-time streaming, Session management...
├── guides/ # Caching, Prompt Engineering, Frameworks, Safety...
└── resources/ # Rate Limits, Billing, Troubleshooting, Release Notes...
api_reference/
├── gemini_api_reference.md # Complete API reference overview
├── api_versions.md # API versioning
└── capabilities/ # Method-level API docs (generateContent, files, etc.)
Use the included filter_docs.py to create a copy of the docs filtered to only your language(s) of choice:
# Copy and filter to Python + REST only (default)
python filter_docs.py --source docs --destination filtered_docs
# Filter for JavaScript only
python filter_docs.py --source docs --destination filtered_docs -l javascript
# Filter in-place (modifies files directly)
python filter_docs.py docs -l python
# Filter a single file
python filter_docs.py docs/core_capabilities/text.md -l goSimply clone the repo and start reading. If you're using this with an AI assistant, point it towards specific files for instant context:
# Example: Find all function calling examples
grep -rn "function_calling" docs/
# Example: Search for generateContent usage
grep -rn "generateContent" docs/Whether you're building a new AI application, integrating Gemini capabilities, or debugging a complex multi-modal pipeline, having the docs at your fingertips (without the browser overhead) is a game changer.
- Fast: No loading states, just text.
- Offline: Works in the air, on a train, or in a bunker.
- Accurate: Sourced directly from the official documentation.
Note: This repository is a community-maintained mirror and is not affiliated with Google.