a2cr/a2cr

A2CR is an open-source local MCP workspace for AI agent handoffs. It lets
Codex, Claude Code, Cursor, and other MCP-capable agents save WorkBaton
checkpoints, store temporary WorkStash notes, coordinate through WorkThreads,
and resume long coding work from a fresh AI window.
Long AI work usually breaks at the handoff. A fresh AI window needs the goal,
current state, decisions, blockers, validation, and next action, but not a whole
noisy transcript. A2CR keeps that handoff state compact, explicit, and safer to
share between sessions.
Use A2CR when you want to:
- restart a long AI coding task from a clean context window
- pass work state between Codex, Claude Code, Cursor, or another MCP client
- keep milestone checkpoints without storing full chat transcripts
- separate compact resume state from optional supporting notes
Japanese overview | MCP setup |
Usage guide | WorkBaton spec |
Local mode spec |
0.1.8 release notes
Directory Status
A2CR is published in the
official MCP Registry
as io.github.a2cr/a2cr-mcp. The local-only 0.1.8 public release is live on
PyPI, GitHub Releases, and the official MCP Registry. See the
distribution inventory for the
current registration and mirror status.
A2CR is also listed and evaluated on the
Glama MCP Registry. Downstream
directories may temporarily show cached metadata after a release; the official
MCP Registry and this repository are the source of truth.
Local Storage Boundary
A2CR now treats the local workspace as the public product boundary. The public
a2cr-mcp wrapper stores WorkBaton, WorkStash, WorkThread, actor, and event
records in a SQLite database on the user's machine. It does not require an A2CR
account, API key, hosted relay, dashboard, or cloud sync path.
The earlier hosted/SaaS relay path is being retired from the public
distribution. New public setup, MCP Registry metadata, and Anthropic Directory
submission work should use the local wrapper only.
Quickstart
Choose the local MCP distribution path that matches your AI client:
| Path | Best for | Distribution | Notes |
|---|---|---|---|
| Python stdio wrapper | Codex, Claude Code, Cursor, generic MCP clients | PyPI package a2cr-mcp | Full public wrapper path for WorkBaton and WorkStash. |
| Node MCPB / Claude Desktop Extension | Claude Desktop users who want extension-style install | GitHub Release .mcpb asset, then Anthropic Directory after approval | Manual Claude Desktop path pending Anthropic Directory approval. No npm install is required for end users. |
Keep the Python wrapper version and Node MCPB compatibility version aligned.
For the current public release, both the Python wrapper and Node MCPB
compatibility version are 0.1.8.
Python wrapper install:
Register A2CR for Codex and verify the local workspace:
Open the local browser dashboard:
a2cr ui binds to 127.0.0.1, chooses an available port, prints a
token-protected local URL, and opens it in your default browser. If the browser
does not open, copy the full printed A2CR_UI_URL, including ?token=..., into
a browser on the same computer. The bare 127.0.0.1: URL is rejected by
design. Keep that terminal running while you use the dashboard. Press Ctrl+C
to stop it.
For a fixed port or a headless/server-style launch:
Use --no-browser when installing on a machine where the terminal cannot or
should not launch a browser; it still prints the full local URL to copy.
Then restart Codex and use the a2cr-local MCP server. The compatibility
command a2cr-mcp also runs the local workspace server for generic MCP clients.
Start with a compact WorkBaton save/resume. Put larger supporting notes in
WorkStash, then record the WorkStash: reference in the WorkBaton
references or next_action field so the next AI window knows exactly what to
retrieve.
Codex-style TOML:
Generic MCP JSON:
After connecting a new AI window, call get_account_limits once, then use
resume_context to continue prior work or save_context to save a new