Argon is a self-hosted, open-source system for Git-like branching and merge workflows in MongoDB, built for AI agents. It is described as giving each agent its own branch of a database so that work can be reviewed, merged, discarded, or undone without touching production.
The product supports isolated branches created in milliseconds, with branches described as a real MongoDB database rather than a copy. It presents a familiar workflow with commands and actions such as creating projects and branches, checking out a branch, viewing history, diffing, merging with preview and apply, restoring to a previous state, and undoing an individual actor’s session. The page also mentions document-level three-way diff, time travel to any state by LSN or timestamp, append-only history, per-actor undo, and reviewable data PRs. Agents can drive the workflow themselves through MCP, and the site says the tool uses an append-only log with pre- and post-images and actor and LSN views.
Argon is framed for AI agent workflows, but the page also references a standard Git-like workflow that database users can recognize. It provides a command-line interface with a Homebrew install command, and the page also shows npm, pip, and SDK references. The site says it is built in the open, property-tested in CI, and that benchmark numbers reproduce with Docker Compose. The licensing and deployment model are stated plainly: open source, MIT licensed, and self-hosted. The page identifies the maker as Argon Labs.
In the Databases (SQL, NoSQL, vector, graph) space, Argon Labs takes a focused approach. It focuses on managing isolated, reviewable database branches for AI agents and developers without risking production data. Argon Labs is an open-source project aimed at developers. Argon Labs is open source under the MIT license. Argon Labs is available on the command line, and it can be self-hosted.
It is developed by Argon Labs, and it first shipped in 2025. The project is developed in the open on GitHub with 166 stars and 42 commits in the last 90 days. Key capabilities include branching, merge requests, and agent workflow.
Summary written by a language model from the project’s public pages.
What PulseGate has recorded for this listing
Closest matches by what these projects do