I replaced a memory app with two Markdown files and a Git repo
Git-native AI memory that works across tools without installing anything.
Open-source team memory layer for AI coding agents — markdown files in git, user→team→org hierarchy, cross-vendor MCP server. Apache-2.0.
Git-based memory you own beats vendor databases — review facts in PRs.
Developers using AI coding agents (Claude Code, Cursor, etc.)
Mem0 · LangChain Memory · Zep
Git-native AI memory that works across tools without installing anything.
Git repo as versioned agent memory — no database, just markdown files.
Issues as versioned Markdown files means you can branch, revert, and diff task history with git.
Markdown files as source of truth means you can always grep your agent's memory.
Issues as Markdown files in your repo means they finally travel with your branches.
Git-verified memory ledger beats vague agent context files.