Murmell is an innovative shared cloud canvas designed to bridge the gap between human development teams and AI agents. Operating as a unified workspace, it allows humans and AI to work together in the same room and on the same repository, complete with live previews and persistent cloud execution that keeps your projects moving even after you close your laptop.

Built to handle collaborative coding at scale, Murmell features a smart file-claiming mechanism where AI agents must claim files before writing to them. This ensures that people and agents can move simultaneously without risking overwritten code. Supporting a diverse lineup of models and tools including Claude Code, Codex, Kimi, and OpenCode—with future additions like OpenClaw and Hermes on the way—Murmell ensures that all completed work safely lands back in git so your projects are always right where you expect them.

Pros of Murmell

Enables real-time collaboration between human teams and multiple AI agents in the same repository.
Smart file-claiming mechanism prevents people and AI agents from overwriting each other's work.
Persistent cloud execution keeps tasks running even after you close your laptop.
Supports multiple AI tools and models like Claude Code, Codex, Kimi, and OpenCode, with more on the horizon.
All completed work automatically lands back in git for easy project tracking.

Cons of Murmell

Limited to the specific AI agents and models currently supported or announced.
Relies entirely on a cloud-based canvas environment which may require constant internet connectivity.
Feature set is currently focused on core repository collaboration and previews, with more features still coming in the future.

Frequently Asked Questions

What is Murmell?

Murmell is a shared cloud canvas where development teams and AI agents can work together in the same room, on the same repository, with built-in previews and persistent cloud processing.

How does Murmell prevent overwriting code between humans and AI?

Murmell utilizes a file-claiming system where AI agents must claim files before they write to them, allowing people and agents to code simultaneously without overwriting each other's changes.