CodeAlmanac – Self-updating wiki for your coding agent (local, Apache)
Auto-updating Markdown wiki from agent chats solves context loss better than vector DBs.

Replaces brittle RAG pipelines with a self-updating wiki that agents can actually read.
Developers building AI agents, technical teams managing documentation
GitBook · Obsidian Publish · Cursor (context features)
I wanted something where:
1. I could use it for myself or I could share it with my teammates.
2. I could upload sources, and the wiki gets created and updated automatically.
3. It is extremely agent-native, because my primary mechanism of interaction would be through my CC or Codex.
So I built Almanac. It takes whatever sources you upload, organizes them, and creates a wiki on top of it. Every time you upload new sources, it reorganizes, restructures, and updates the wiki.
The agent interacts with the wiki through a CLI. Similar to Vercel, you can link a folder locally to a wiki, and your agent will then automatically have context of it.
I also wanted to make it easy to develop on this platform. I think wikis are a much better way of giving your agent access to quality information compared to RAG. So there is also a Python SDK as well as an MCP that you can use to build on top of it.
The project is still early, and I would love any feedback or feature requests that you have.
Btw we've also made an open-source version of Almanac for code called CodeAlmanac(https://github.com/AlmanacCode/codealmanac)
Auto-updating Markdown wiki from agent chats solves context loss better than vector DBs.
Packages Karpathy's wiki pattern into Claude Code plugins for persistent project context.
Automates wiki bookkeeping so LLMs update cross-references while you just drop in raw sources.
Deterministic merge reconciliation for parallel agent wiki edits.
AI docs CLI inspired by OpenWiki in a crowded agent documentation space.
Auto-ingests wikis to fix agent SQL accuracy when Claude Code hallucinates metrics.