Open Source DeepWiki: AI-Powered Wiki Generator for GitHub/Gitlab Repos (github.com)

DeepWiki (deepwiki-open) is an open-source tool that auto-generates interactive, navigable wikis for any GitHub, GitLab, or BitBucket repository. Point it at a repo (including private ones via personal access tokens) and it clones the code, analyzes structure, creates embeddings, generates context-aware documentation, auto-produces Mermaid diagrams, and organizes everything into a browsable wiki with an Ask feature that uses RAG-powered Q&A and a DeepResearch multi‑turn investigation flow. The project includes a FastAPI backend (rag, data pipeline), a Next.js frontend, and Docker Compose support for one‑command deployment. Technically notable: DeepWiki builds vector embeddings and a retriever pipeline for accurate code search and RAG, supports multiple LLM providers (Google Gemini by default, OpenAI gpt-5-nano default, OpenRouter, Azure OpenAI, and local Ollama models), and exposes generator/embedder/repo JSON configs to tune models, chunking, and repo filters. It also supports enterprise workflows via OPENAI_BASE_URL for private endpoints, swap-in OpenAI-compatible embedders, and local model hosting. Operational features include configurable logging, secure log path enforcement, and environment-driven configuration. For developers and engineering orgs, DeepWiki streamlines onboarding, documentation maintenance, and code discovery while offering flexible model/provider choices and private-repo support.