DeepWiki Open is an open-source, AI-powered wiki generator for code repositories, with more than 18,200 stars on GitHub. Point it at a GitHub, GitLab, or Bitbucket repository and it analyzes the codebase and produces a structured, navigable wiki — architecture overviews, module documentation, and diagrams — plus an interactive assistant you can ask questions about the code. It is the community open-source take on the DeepWiki idea, and it remains in active development in 2026.
DeepWiki Open clones and indexes a repository, builds a knowledge graph of how files and components relate, and uses retrieval-augmented generation grounded in the actual source to write its documentation and answer questions. It supports multiple model providers — OpenAI, Gemini, OpenRouter, and local models through Ollama — so teams can balance quality, cost, and privacy. Generated wikis include visual diagrams of system structure, and the whole application can be self-hosted with Docker, keeping proprietary code inside your own infrastructure while still getting an always-up-to-date, queryable map of the project.
DeepWiki Open is completely free and open source under the MIT license. You pay only for any model API usage, or nothing at all when running local models with Ollama.
DeepWiki Open is for engineering teams, open-source maintainers, and new hires who need to understand an unfamiliar codebase fast without reading every file by hand. It is equally handy for keeping internal documentation in sync with a fast-moving repository, since the wiki can be regenerated as the code evolves.