DeepWiki, AutoWiki and OpenWiki point to a new pattern for agent context
garrytan · x · 2026-07-23
A thread argues that the emerging “LLM wiki” pattern is becoming a standard way to feed agents better context: compile docs once, keep them updated, and have the agent read the generated pages instead of raw sources.
It cites several implementations:
- Cognition launched DeepWiki, a generated wiki for every public GitHub repo, with Devin reading that instead of raw code.
- Factory built AutoWiki, which treats docs as a build artifact and regenerates them in CI on every push.
- LangChain open-sourced OpenWiki, extending the idea from repos to a person’s working life across Gmail, Notion, and Git.
- Garry Tan open-sourced GBrain, a personal-scale version built from files in Git.
The common idea is to compile once at ingest, maintain the pages, and route agents to the curated wiki rather than the original noisy sources.
More from coding & agent
- Cognition's SWE-2 uses a KKT duality argument in RL to shift the effort Pareto curve — YouJiacheng · 2026-09-11
- First-ever Three.js Conference lands in Paris, with a panel on AI-shortened design workflows — OdinLovis · 2026-09-11
- Data engineering, not agent frameworks, is the real bottleneck for enterprise AI agents — dhruv2038 · 2026-09-11
- RTK Terminal Compression Cuts Tokens but Leaves Your AI Coding Bill Unchanged — Bartaseth · 2026-09-11
- GPT-6 Astra beats Factorio with enemies in 44 in-game hours at ~$4,500 API cost — liminal_bardo · 2026-09-11
- Investment Analyst Asks How to Build a Claude-Based Diligence Agent Stack — Careless_Tie2286 · 2026-09-11