Features
Version control for your MongoDB data
Lightweight metadata branching, time travel, and restore, built on a deterministic write-ahead log. Open source and self-hosted.
/01
Millisecond branching
Metadata branches share history. Physical checkout creates a MongoDB database and has dataset-dependent time and storage costs.
/02
Time-travel queries
Inspect retained historical states, addressed by log position (LSN) or timestamp. Pins protect their referenced history.
/03
Deterministic by design
Same history, same state — byte for byte, property-tested in CI.
/04
Agent & eval workflows
Managed TTL sandboxes, branch-scoped actor undo, dataset pins, and a LangGraph checkpointer (argon-agents).
/05
Four surfaces, one engine
CLI for humans · MCP for agents (13 tools) · REST for SDKs · wire proxy for stable URIs.
/06
Honest engineering
Every number links to a benchmark you can run with docker compose up. What we can't back, we don't publish.
What to expect
Branch vs checkout
A branch is lightweight metadata. Checkout materializes a real MongoDB database; readiness, disk use and first-query costs depend on the dataset.
Isolation and merge
Experiments write to a separate branch database. Applying a reviewed merge explicitly changes the target branch. Use MongoDB credentials and network controls for access isolation.
Writer attribution
Native driver writes use the actor configured for the branch or run. Separate agents need separate branches; Argon does not identify individual clients sharing one connection.
Capture and undo
A healthy capture process records document changes. Undo requires complete images and retained history. Missing images or unsupported drop/rename operations mark history incomplete and prevent unsafe restoration.
Retention and pins
Time travel is available within retained history. Pins protect the data they reference while the pin exists; keep backups and choose retention for your workload.
Sandbox lifecycle
REST and MCP manage capture and TTL cleanup while the service runs. Standalone CLI sandbox creation requires a running watch process and scheduled sandbox sweep.
Anonymous demo
The hosted demo provides session-scoped sample data and review tools. Native MongoDB connections and physical sandboxes are available in your local deployment.
Everything is a view over one append-only log: branches are pointers into it, time-travel replays it (deterministically, property-tested), and checkout materializes it into a real MongoDB database. The deep dives: