Murmell

Product Introduction

1. Definition: Murmell is a cloud-based collaborative development environment (CDE) and multi-agent orchestration platform. It functions as a shared, persistent workspace where human developers and multiple AI coding agents (like Claude Code, Codex, Kimi, OpenCode) can concurrently operate on the same code repository in real-time.
2. Core Value Proposition: Murmell exists to eliminate the friction and isolation inherent in using AI coding assistants at scale. Its primary value is enabling true collaborative AI development by providing a single source of truth, preventing agent conflicts, and allowing work to continue autonomously in the cloud, independent of a developer's local machine state.

Main Features

1. Shared Cloud Canvas & Persistent Workspace: Murmell hosts a dedicated virtual machine for each project, providing a browser-accessible development environment. This canvas persists 24/7, meaning agents continue executing tasks even after a user closes their browser. All work is automatically synchronized to a dedicated private GitHub repository, ensuring code is never lost and version control is maintained.
2. Multi-Agent Concurrency with File Locking: The platform's core technical innovation is a file-level locking mechanism. Before an AI agent writes to a file, it must claim an exclusive lock on that file path. This prevents race conditions and overwrites between multiple agents (or humans) operating simultaneously. The system enforces "one holder per path," serializing edits at the infrastructure level rather than relying on agent prompts.
3. Real-Time Collaborative Interface: The Murmell canvas supports multiple live viewers and participants. Every user on the shared link sees real-time cursor positions, active user names, terminal output, and agent activity without manual refresh. This enables pair programming and oversight with AI agents, transforming them into visible team members.
4. Bring-Your-Own-Agent (BYOA) Model: Murmell does not resell AI model access. Instead, it launches the command-line interfaces of supported AI coding tools (Claude Code, Codex, etc.) on its cloud machines, using the API keys or subscriptions provided and managed by the user. This decouples the collaboration infrastructure cost from the variable AI inference costs.
5. Integrated Agent Orchestrator: The platform includes a native orchestration agent. Users can issue high-level natural language commands (e.g., "spawn three agents on the arcade and wire the preview"), and Murmell will automatically provision the specified agents, assign them files from the repository, and coordinate the initial setup.

Problems Solved

1. Pain Point: Agent Isolation and Merge Hell. Using multiple AI agents locally or in separate sessions leads to fragmented work, conflicting edits, and manual, time-consuming merge operations. The "Friday of merges" scenario is a direct consequence of uncoordinated, parallel AI work.
2. Pain Point: Lack of Persistent AI Execution. AI coding sessions are typically tied to a local IDE and laptop. Closing the laptop terminates all agent processes, halting long-running tasks, background refactoring, or autonomous feature development.
3. Target Audience: Development Teams and Tech Leads who are scaling their use of AI assistants beyond individual "copilot" use. Startup and Product Teams building rapidly with small teams. Solo Developers and Indie Hackers who want to leverage multiple AI "teammates" to increase output and manage complex projects.
4. Use Cases: Complex Feature Development requiring parallel work on interconnected files (e.g., frontend component, backend API, and database schema). Continuous Code Refactoring & Tech Debt Reduction with dedicated agents running overnight. Live Collaborative Development Sessions where team members and AI agents brainstorm and build together in real-time. Onboarding & Code Review, where new team members can observe AI agents and colleagues working in a shared context.

Unique Advantages

1. Differentiation: Unlike single-agent IDE plugins (GitHub Copilot, Cursor) or isolated cloud IDEs, Murmell is architected for multi-agent concurrency as a first-class feature. It competes with manual process orchestration, not with individual AI coding tools. Its pricing model is also distinct, charging for the persistent collaboration environment and compute, not for AI tokens.
2. Key Innovation: The agent-aware file system lock is the critical technical differentiator. By managing resource contention at the OS/workspace level, it allows unmodified, off-the-shelf AI agent CLIs to work together safely. This turns a "traffic jam" of five agents into a coordinated assembly line without requiring custom agent logic or post-hoc merge tools.

Frequently Asked Questions (FAQ)

1. How does Murmell prevent AI agents from overwriting each other's code? Murmell implements a mandatory file-locking system at the workspace level. Before any agent (or user) can write to a file, it must successfully claim a lock on that specific file path. If a file is locked, other agents are queued and wait until it is released, eliminating merge conflicts at the source.
2. Where is my code stored, and is it secure? Your code resides in two primary locations: a private GitHub repository that you own, created automatically with the project, and the ephemeral workspace on Murmell's isolated project-specific virtual machine. The cloud workspace is disposable and contains no persistent database or decryption keys, with all changes pushed to your private Git repo.
3. Do I pay Murmell for AI model usage (tokens)? No. Murmell operates on a bring-your-own-keys (BYOK) model. You provide and manage the API keys or subscriptions for Claude, Codex, Kimi, etc. Murmell's subscription fee covers the cloud compute, collaboration platform, and orchestration infrastructure, not AI inference costs.
4. Can I use Murmell for free? Watching a live, shared canvas is free and requires no account. For creating and running your own projects, all paid plans (Solo, Pro, Builder) start with a 14-day free trial. Pro and Builder tiers also include promotional credit for specific AI providers.
5. What happens if the Murmell service goes down? Since your primary code repository is a private GitHub repo, your source code remains safe and accessible. The cloud workspace is stateless relative to your code. Service downtime would affect the active collaborative session and agent execution, but not your code assets, which are version-controlled externally.