MongoDB sandboxes
for AI agents
Git for MongoDB underneath.
Give every agent run its own real MongoDB database, forked from your data. Review what it changed, merge what works, and discard the rest.
Open source under MIT. Self-hosted, on your own MongoDB.
orders / order-1
Illustrationmain
- price
- 49
Unchanged
planner
- price
- 49
Same baseline
Start from the same data.
Give the planner agent its own branch of orders. Both start with the same document.
Illustrated workflow. Run the real engine locally. How it works
npm install -g argonctl@2.9.0 --allow-scripts=argonctlInstalls the argon CLI with npm 11.16+ and a supported Node.js version. Allows only argonctl's binary installer.
Published release: v2.9.0
One CLI named argon. It runs the engine, the local console, the MCP server and the wire proxy.
Git for MongoDB, command by command
Every sandbox is a branch, so the workflow is the one you know from Git.
- Give every run its own database
argon sandbox create - Review changes as a diff
argon diff - Merge the plan you reviewed
argon merge preview - Read or restore an earlier state
argon time-travel query - Undo one agent's writes
argon undo - Start every eval from the same data
argon pin create - Start from a database you already run
argon import databaseTested on MongoDB 7.0 replica sets only: not on Atlas or with access control. - Write a reviewed change back
argon merge previewFrom the CLI only. Tested on MongoDB 7.0 replica sets only: not on Atlas or with access control.
Use the tools you already have
An agent branch checks out as a real MongoDB database. Keep your usual driver; use Argon to manage its history and review its work.
- Set up MCP
Connect an agent
Give Claude Code or Cursor database tools through MCP.
- Use the Python SDK
Build in Python
Run the review workflow with the SDK, then add it to your agent.
- Call the REST API
Use your own stack
Create sandboxes and inspect changes from any HTTP client.
- See the boxd integration
Pair with a code sandbox
Run agent-written code on an isolated boxd machine against an Argon sandbox.
Verify it yourself
The two-agent example checks the starting data, a real conflict, the accepted change, and the restored result.
Review your first change
Start with a sample order. See what the agent changed.