Product

One control layer for your AI engineering crew

MUON sits between your engineering mission and the coding agents doing the work. It assigns bounded responsibilities, keeps context moving, tracks progress, and returns consequential decisions to your team.

See how the loop works

The platform contract

MUON runs the crew. Your team sets the mission, approves sensitive actions, and decides what ships.

Working surfaces

One mission record, however your team works.

Full workspace

Desktop

Mission chat, Crew, shared memory, approvals, changes, governed review, agent sessions, native terminals, and audit activity share one workspace-bound interface.

Complete control path

CLI

Onboard providers, run tasks, apply workflows, inspect status, review approvals, request a second opinion, run checks, and file merge approval from scripts or a terminal.

Operational parity

TUI

Observe the same jobs, gates, memory, workflow state, and panic-stop outcome in a keyboard-first interface designed for narrow terminal layouts.

Agent stack

Clear roles, governed by MUON.

Claude CodeCodexCursorOpenCode
Claude CodeDispatch-ready

MUON can assign bounded implementation, investigation, repair, or review work and create governed child jobs with enforced lineage and limits.

CodexDispatch-ready

MUON can assign bounded implementation, investigation, repair, or review work and create governed child jobs with enforced lineage and limits.

CursorDispatch-ready (scout/review)

MUON dispatches Cursor as a managed read-only lane for scout, reviewer, QA, and architect roles under deny-first permissions. It is not an implementer seat.

OpenCodeDispatch-ready (scout)

MUON runs OpenCode as a scout-only managed lane under a deny-first permission table, with the governed MUON brain injected when the vendor accepts the MCP block.