Agent Wikis

Agent Wikis compiles messy, fast-moving sources into maintained, structured wikis. It’s for teams building AI agents that need reliable, niche knowledge, and for humans who want a clearer place to browse that information.

• Curated wikis built from sources like repos, changelogs, video transcripts, and community Q&A
• Markdown pages designed for direct reading (concepts, entities, syntheses, release notes)
• llms.txt entry point so agents can start from a simple, agent-readable manifest
• Scope and freshness metadata attached to each wiki
• RAG search over maintained content (compiled pages, not raw chunks)

The key idea is that accuracy depends on having a curated knowledge base, not just pulling chunks from the live web. Agent Wikis focuses on compiling sources into structured pages and updating them as sources change, with a declared scope and freshness signals.

You can browse public wikis without an account, using the same raw Markdown and llms.txt interfaces agents can read. The site includes multiple categories such as AI Agents, Inference & Serving, Image & Video Gen, Blockchain & DeFi, Trading & Markets, Marketing & Ads, and Developer Tools.

Public wikis are free for everyone. There are paid tiers for larger, deeper “XL” editions and for teams that want to claim a wiki for their product with input on scope, sources, and freshness, plus an eval report for how accurately agents answer from it.