Run parallel, isolated agents on your infrastructure.
Where developers scale their leverage with parallel coding agents and platform teams maintain enterprise-grade boundaries, governance, and auditability.
Scale your leverage with parallel agents.
Run multiple AI coding agents simultaneously, each in its own isolated workspace. Review plans before execution, delegate sub-tasks, and give feedback to keep agents on track.
- Parallel Agents — Run multiple agents simultaneously in isolated environments, each with its own Git branch and chat history.
- Spec-Driven Development — Have the agent write a detailed plan file for your review and approval before any code changes happen.
- Sub-Agent Delegation — Spawn background agents to explore codebases, investigate problems, or try multiple solutions.
- Secure Message Sharing — Share workspace conversations with teammates while keeping sensitive data protected.
- Built-In Code Review — See exactly what changed with diff views, give code feedback, and track Git status across all workspaces.
- Custom Agents — Define your own agents with Markdown files to control tools, prompts, and behaviors.
Deploy agents with enterprise-grade governance.
Run Mux in Coder Workspaces on your infrastructure for full control over security, compliance, and access. Enforce boundaries, audit activity, and integrate with your existing identity stack.
- Coder Workspace Integration — Run agents inside governed, self-hosted infrastructure with centralized provisioning and policies.
- Audit Logging — Track all agent activity with full observability — see which agents did what, when, and in which workspace.
- Model Governance — Control which LLM models developers can use, including private endpoints and enterprise API keys.
- Agent Firewall — Enforce what agents can access with process-level firewalls that restrict domains, commands, and filesystems.
- Enterprise Authentication — Integrate with your existing identity provider via OIDC/SAML for secure, centralized access control.
- Open Source & Self-Hosted — Fully inspectable AGPL codebase that runs on your infrastructure — no data leaves your environment.
Run Mux anywhere.
Use Mux on macOS, Linux, or Windows through the desktop or browser app, and scale up by connecting to remote compute when your workloads grow.
Bring any model.
Pick the best model for the task. Mux makes it easy to switch between state-of-the-art models without forcing developers to swivel chair between tools.
Audit every agent action.
Pair Mux with Coder’s governance controls to run AI agents on enterprise infrastructure with full activity logs, policy enforcement, and the visibility security teams expect.
Open source and always free.
Inspect, modify, and extend Mux as you like. The core experience is fully open source and free for individual developers and teams.