Open-source LLM tooling, reviewed from the source code
10 public projects in 5 categories. Each project has a DeepWiki-style overview. Each category asks the same "how is X implemented?" questions of every project, with answers that cite the code at a pinned commit.
Open-source personal assistants
Self-hostable AI assistants that act on your mail, calendar, docs and chat on your behalf.
2 projects · 6 questions →Browser & computer control
Frameworks that let an LLM drive a browser (or a whole desktop/phone) to complete tasks.
2 projects · 6 questions →API layer & connectors
Platforms that give agents and apps authenticated access to hundreds of third-party APIs.
2 projects · 6 questions →AI web scraping
Crawlers, extractors and stealth browsers that turn web pages into LLM-ready data.
2 projects · 6 questions →Open-source DeepWiki
Self-hosted generators that turn a code repository into a browsable, AI-written wiki.
2 projects · 6 questions →Unlike vendor docs or auto-generated wikis, every claim here is checked against the code. Each category also has a one-page comparison: Assistants · Browser control · Connectors · AI scraping · DeepWiki clones.
How this site is made
Every repository is cloned at a pinned commit. Two open-source DeepWiki generators (deepwiki-open and OpenDeepWiki) write a machine wiki for each one. A code-reading agent (Claude Code running DeepSeek V4 Flash) answers each category question with file-and-line citations. Each citation is checked mechanically. An editor model writes the overviews and the cross-project verdicts. Read the full method →
For AI agents
Every page has a markdown twin (append index.md, or send Accept: text/markdown). There is also a read-only JSON API (OpenAPI), an MCP server at /mcp, llms.txt, llms-full.txt and AGENTS.md. Crawling, AI answers and AI training are all welcome.