Murmell is a browser-based canvas where several AI coding agents work one project directory, with live terminals and file reservation to stop collisions.
Overview #
Murmell #
Murmell is a shared cloud canvas for AI coding agents. Murmell runs Codex, Claude and Kimi agents inside one project directory in the browser, shows each of them working in a live terminal, and uses file reservation so two agents do not edit the same file at once. Teams watch the same canvas together, which turns agent work into something a group can follow rather than something happening in one developer's terminal. You connect your own agent subscriptions and API keys; Murmell provides the cloud infrastructure and the coordination layer.
Production credibility: Murmell launched publicly in August 2026 and publishes its product openly: a browser-based shared canvas, real-time terminals per agent, file reservation to prevent conflicts, and a bring-your-own-keys model where users supply their existing agent subscriptions. It is early — the company is small and pricing is not published on the site — so it is best evaluated as a young product with a specific and well-defined problem rather than as an established platform.
Key Features
- Browser-based canvas shared by a team in real time
- Multiple agents — Codex, Claude, Kimi — in one project directory
- Live terminal view for each running agent
- File reservation to prevent two agents editing the same file
- Bring your own agent subscriptions and API keys
- Cloud infrastructure managed by Murmell rather than run locally
Ideal Use Case
Small teams that want to run several agents on one codebase and watch the work together — pairing on agent supervision rather than on typing. It suits groups already sharing agent subscriptions who want the coordination handled centrally, and it is a cloud alternative for teams that do not want a local Mac-only setup.
How Murmell differentiates
Conductor solves the parallel-agent problem locally on a Mac; Murmell solves it in the browser for a group. The shared canvas and file reservation are the specific contributions: instead of each developer running their own agents privately, the team sees one workspace and the system arbitrates who edits what. The trade-offs are maturity and trust — it is a new product, pricing is not public, and it needs your agent credentials to run.
FAQ
Q: What is Murmell? A: Murmell is a browser-based shared canvas where multiple AI coding agents work in one project directory, with a live terminal for each agent and file reservation to prevent conflicts.
Q: Which agents does Murmell support? A: Murmell runs agents including Codex, Claude and Kimi. You connect your own agent subscriptions and API keys rather than buying model access from Murmell.
Q: How much does Murmell cost? A: Pricing is not published on the site. Because you bring your own agent keys, model usage is billed by your existing providers.
Q: How does Murmell prevent agents clashing? A: Through file reservation: an agent claims the files it is editing, so two agents cannot modify the same file at the same time in the shared directory.
Q: Is Murmell a local or cloud tool? A: Cloud. It runs in the browser with Murmell managing the infrastructure, which is the main difference from local parallel-agent tools such as Conductor.
tl;dr
Murmell is a browser-based shared canvas for running Codex, Claude and Kimi agents on one project directory, with live terminals, file reservation and team collaboration. Bring your own agent keys; pricing is not published.