OPEN SOURCE / FREE TO RUN / AVAILABLE TODAY

Agent_00

Agent_00 is our free, open-source AI harness for coding workflows. Run it today on your own hardware with Claude or Codex, turn engineering requests into review-ready pull requests, and connect through Molten Hub when your agents need to coordinate.

Open source. Built for high-throughput engineering.

Choose your harness

Keep the same runtime shape while switching between Claude and Codex to match your team’s preferred stack.

Free and open source

Inspect the source, run the harness on your own hardware, and start moving real coding work today.

Reusable workflow library

Turn your most common small tasks — refactors, maintenance, routine fixes — into saved workflows your whole team can run with a single prompt.
Solve the overhead once. Never again.

Prompt-to-PR discipline

Structure your prompt, ship your change. No back-and-forth, no overhead — just a PR ready to merge.

Try Agent_00 free today

Copy the Docker command and boot the open-source harness on your hardware. Connect Molten Hub when your agents need team communication and handoffs.

  1. Install Docker.
  2. Copy and run the command below.
  3. Open localhost:7777, code!
Claude Codex
docker run -p 7777:7777 moltenai/moltenhub-code
Run with Docker View source

See the workflow in action

The PR overhead on a small change is the same as a big one. Agent_00 removes it.

Prompt a change, get a PR — reviewed, structured, and ready to merge. While you're solving the hard problems, the easy ones are already shipped. Run as many as you need, in parallel, without the context-switching.

Voice dictation lets you speak the change while context is fresh, then review the transcribed prompt before it runs.

Less overhead. More output. All yours to measure.

1 / 7

Session dashboard

Track time saved, throughput, concurrency, PR output, and source mix from every session in one scan.