strategy

target problem

every team using ai tools re-explains context across chats, agents, and team members. teams working with ai agents at scale need a single source of truth that every agent can read. existing solutions (saas products, wikis, manual files) are either hosted (data leaves your control), unstructured, or lack a standard ai protocol.

our approach

we make the organisation's context a git repo with an mcp server on top. the repo is the backend — it stores everything. git provides version history, branching, pr review, and access control. the mcp server gives every ai agent a standard protocol to read, search, and update it.

three principles:

1. open source, forkable — one click, and your team has its own org memory.
2. git-native — no database. the repo is the database. all changes go through prs.
3. mcp protocol — any mcp-compatible ai tool connects directly.

who it's for

engineering teams, startups, and open-source projects that use ai coding tools and need persistent organisational context.

key metrics

• forks of the template repo
• github stars
• npm downloads of the mcp server

tracks

track	what	priority
server	mcp server with read/write/search/pr tools	p0
template	the organisation.md file with all sections	p0
docs	readme, quickstart, agent prompts	p0
dx	easy setup, clear error messages	p1
extensions	archive pruning, multi-file updates	p2

milestones

1. v1.0 — forkable repo with mcp server, template, docs, github pages
2. community adoption — real teams using it, contributing improvements
3. extension system — pluggable context sources and exports