Automated repository documentation
A specialized AI agent analyzes repository source code and generates natural-language summaries, so readers can get oriented without manually writing or maintaining docs.
Code Wiki is an AI-powered documentation hub for code repositories with summaries, architecture diagrams, and repo Q&A to help developers navigate codebases faster.
Code Wiki is an AI-powered documentation hub for code repositories. It analyzes source code and generates natural-language summaries, architectural diagrams, and other repository insights to help developers understand a codebase faster.
The product also includes a conversational “Ask about this repo” experience that uses the generated wiki content as its primary context. That makes it useful both for initial onboarding and for follow-up questions about structure, behavior, and repository details.
A specialized AI agent analyzes repository source code and generates natural-language summaries, so readers can get oriented without manually writing or maintaining docs.
The wiki includes architectural diagrams that show relationships, hierarchies, and dependencies in the codebase, with interactive viewing for zooming and panning.
A conversational experience lets users ask questions about a repository using the wiki content as the main context source.
Search lets you look up a repository by name from the homepage, and a direct URL format also points from a repository page to its wiki.
Wikis refresh periodically rather than staying static, so the generated documentation is updated over time as the source repository changes.
Public wiki pages are read-only and do not accept direct edits or pull requests, which keeps the generated documentation aligned to the source repository.
A developer joining a new repository can use the generated summaries and diagrams to understand the structure of the codebase before opening the source in detail.
A team can use the wiki as a reference when they need quick answers about a repository instead of searching through files manually.
Someone evaluating a public open-source project can browse the wiki to get a higher-level view of architecture and key components before reading the full code.
A repository owner can rely on periodically refreshed documentation to keep generated content aligned with source changes without maintaining a separate manual docs process.
Code Wiki is an AI-powered documentation hub for a code repository. It analyzes source code to generate summaries, architectural diagrams, and other repository insights, and it supports a chat experience for asking questions about the repo.
Code Wiki is built on Google’s Gemini models. The FAQ says a specialized AI agent uses Gemini to analyze source code and generate the wiki content, and the chat experience is also powered by Gemini using the wiki as context.
Public Code Wikis for open-source repositories on codewiki.google are free.
Code Wiki currently supports repositories hosted on GitHub only. The FAQ says support for other platforms such as GitLab and Bitbucket is planned for the future.
Wikis are updated periodically, with refresh timing varying by project based on popularity, usage, and other factors. The FAQ also notes that the last generated date appears at the bottom left of the page.