Egregore CLI
Shared team memory for people and AI agents in Git: decisions and handoffs live in Markdown, reachable from Claude Code and Codex.
- 293 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- Creates GitHub repositories, syncs them and runs hooks on every session.
- Telemetry on by default records local identifiers and metrics.
Install
- Works in
Manual install
npx create-egregore@latestThe command from the Start an Egregore section. After setup run `egregore` in a new terminal.
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Creates GitHub repositories, syncs them and runs hooks on every session.
- Telemetry on by default records local identifiers and metrics.
This is third-party code. Review the repository files before installing.
What it does
Egregore is a CLI with a set of skills and hooks that builds shared organizational memory on top of a separate Git repository. Decisions, findings and handoffs are written to Markdown files and linked into the workspace as `memory/`. A teammate in another session, even with a different agent, finds the relevant decision with its sources. Setup runs through `npx create-egregore@latest`: the wizard guides GitHub sign-in, the instance name and the agent choice. The `egregore` command starts a session with memory sync. Working commands include `handoff`, `reflect`, `deep-reflect`, `quest`, `ask`, `activity`, `dashboard`, `todo`, `save` and `invite`. In Codex they are invoked as `$handoff`, `$reflect`.
Who it is for. Teams where several people work through Claude Code and Codex and lose context when handing tasks between sessions.
Good fit when
- One developer finishes the day and a teammate with a different agent continues the task tomorrow.
- In a month you need to recall why a particular onboarding decision was made.
- The team needs a log of decisions and open questions in plain Markdown files under Git control.
- A new member is invited by GitHub username and immediately gets the memory and project repositories.
Not a fit when
- You work solo and a notes file in the repository is enough.
- Extra Git repositories and GitHub sign-in are not allowed.
- Telemetry cannot stay on by default and manual configuration is unacceptable.
Example request
Find the handoff about the redesign and explain why that approach was chosen, with sources.Limitations
Node.js 22 or newer, Git, jq, Bash and Python 3.11 or newer are required, plus GitHub CLI for sign-in. Semantic search models are fetched during setup, so keyword search comes first. Local memory and search work offline, while GitHub sync and hosted model providers need connections. Telemetry is on by default: it records identifiers and metrics without prompts or code, uploads nothing by itself and is disabled with `EGREGORE_NO_TELEMETRY=1` in `.env`. Egregore Connect, optional hosted services for membership administration and notifications, is described as optional in the README.
How to disable. Delete the Egregore repositories and the local workspace created by setup, then remove `egregore` from PATH. Revoke teammates' access in the GitHub repository settings.
FAQ
Where is the memory stored?
In Markdown files of a separate Git repository linked into the workspace as `memory/`. The organization owns that repository and the decision history stays in Git.
How do I join an existing Egregore?
Ask the owner to invite your GitHub username and run `npx create-egregore@latest join <owner>/<egregore-repo>`. Setup brings in the memory and project repositories.
Can I switch between agents?
Yes. Claude Code, Codex, Pi and Prime Agent are supported. A handoff started in one agent continues in another through the same memory.
What about telemetry?
It is on by default and excludes prompts, code and file contents. Disable it with `EGREGORE_NO_TELEMETRY=1` in `.env`.
What do I need to install it?
Node.js 22 or newer with npm, Git, jq, Bash and Python 3.11 or newer. The setup wizard offers to install GitHub CLI itself.
Discussion
No questions or reviews yet.
Related
CLI toolsMemory for coding agents in plain Markdown: the agent keeps specs, decisions and lessons itself, and each new session starts from them. No vector database
Decision memory in git: AI agent sessions are saved at commit time, and an agent can recall the reasons and rejected options
Persistent agent memory in a folder of linked markdown notes, so Claude Code remembers a project across sessions
A CLI for AI agents sharing a repository: a task board, a spec and acceptance by a second agent against a specific commit
Claude Code Memory Setup: Obsidian-based memory
Claude Code Memory Setup
A guide with ready files: Claude Code keeps decisions between sessions through an Obsidian vault and /save and /resume commands
One config for Claude Code and Codex in every repo: shared rules, skill packs, second-model review and a destructive-command guard
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| Egregore (this card) | no | 293 | ||
| Claude-Mem | no | 113,534 | a plugin that stores one user's session memory; Egregore keeps shared team memory in a Git repository | |
| PLUR | no | not required | 312 | shared agent memory in YAML files with user corrections; Egregore adds handoffs, questions for teammates and invitations |
README in short
The open-source runtime includes memory, local search, coordination workflows and adapters for four agents. Sources live in `bin/`, `.claude/`, `.codex/`, `.pi/` and `.prime/`.