Mission controlfor Claude Code
Orchestrate parallel Claude sessions, review every diff, track costs, and ship PRs — all from one place.
Powered by Claude
Anthropic Agent SDK
Built-in Code Review
Quick, Deep, Security
Native Desktop
macOS, Windows, Linux
100% Open Source
GPL-3.0 License
macOS · Windows · Linux
See what your agents are building
The hardest part of running parallel agents isn't starting them — it's knowing what they're doing. ChatML makes every session visible.
All your agents, one dashboard
See every running session at a glance — status, progress, and what each agent is working on. No more tab-switching.
Every change, reviewed
Built-in diff viewer with inline comments. Run Quick, Deep, or Security reviews before any code gets merged.
Know what you're spending
Real-time cost and token tracking per session. Set budgets so agents don't run away with your API credits.
Everything you need to ship faster
Scope tasks, delegate to agents in isolated worktrees, and review their work async.
Isolated Git Worktrees
Each AI session runs in its own git worktree. Work on multiple features simultaneously without context switching or branch conflicts.
Quick, Deep & Security Reviews
Three review depths for every situation. Quick for fast checks, Deep for thorough analysis, Security for vulnerability scanning.
Visual Code Changes
Review every change with split or unified diff views. Syntax highlighting for all languages with line-by-line comparison.
Integrated Terminal
Run commands without leaving the app. A built-in terminal keeps your workflow in one place — no context switching between windows.
Command Menu for Power Users
Access any action instantly with keyboard shortcuts and a quick command menu. Navigate, switch modes, and execute commands without touching the mouse.
Streamlined Pull Requests
Create, review, and manage pull requests directly from the app. From code to merged in a single workflow.
And so much more
Every detail crafted for professional developer workflows.
Linear Integration
Link Claude sessions directly to Linear issues. Status updates, branch names, and session context sync automatically.
Skill Library
Browse and install skills to give Claude new capabilities. Development, documentation, security, and version control — one click to enable.
Extended Thinking
See Claude's reasoning unfold step by step before it writes a line of code. Understand why, not just what.
Cost & Token Tracking
Per-session cost monitoring with budget limits. Track token spend across all running agents.
Plan Mode & Approval
Claude proposes a plan before touching any code. Approve, edit, or reject — you stay in control.
Sub-agent Spawning
Agents delegate subtasks to child agents automatically. Parallel execution, zero manual coordination.
Session Dashboard
All running sessions in one view. Status, token usage, and progress at a glance.
File Browser & Tabs
Browse project files and track what Claude is reading or writing. Multi-tab editing keeps context intact.
Command Palette
Every action reachable by keyboard. Search sessions, run reviews, switch workspaces — all without the mouse.
Multi-workspace Support
Add and switch between multiple git repositories. Each workspace keeps its own sessions and worktree history.
MCP Server Support
Connect any Model Context Protocol server to extend Claude's capabilities. Databases, APIs, and custom tools — all pluggable.
Activity Notifications
Get notified when agents complete, hit errors, or need approval. Native alerts so you can step away and stay informed.
Built in the open. Backed by the community.
ChatML is 100% open source under GPL-3.0. Every commit, every decision, every line of code — transparent and community-driven.
$ cat stack.toml
[core]
desktop = "Tauri"
backend = "Rust"
api = "Golang"
ui = "Next.js"
license = "GPL-3.0"
We build ChatML in public because developer tools should be transparent. Follow our progress, contribute to the codebase, or just watch how we ship.
Star us on GitHub
Join the developers following our progress. Every star helps.
Follow the build
Daily updates, behind-the-scenes decisions, and release announcements.
From idea to shipped code
Three steps to parallel AI development.
Delegate
Describe what you want built. Spin up 3, 5, or 10 parallel Claude sessions — each one working independently on its own task.
Monitor
Watch every agent in real-time from one dashboard. See diffs as they're written, track costs per session, and catch issues early.
Review & ship
Run multi-depth code reviews — Quick, Deep, or Security — approve changes with inline comments, and create PRs without leaving ChatML.
CLI is powerful. ChatML adds the cockpit.
Claude Code handles the coding. ChatML gives you visibility, review, and control across every session.
Agent visibility
Dashboard showing all sessions, diffs, and progress in real-time
Check each terminal tab individually
Code review
Built-in 3-depth review (Quick, Deep, Security) with inline comments
Manual git diff or external review tools
Cost tracking
Real-time per-session cost and token monitoring with budgets
Basic token count, no per-session breakdown
PR workflow
Create, review, and approve PRs without leaving the app
Separate gh CLI commands in another terminal
Session management
Unified UI to spawn, monitor, and compare all sessions at once
Separate terminal windows or tmux panes
Project management
Linear integration with automatic status updates and branch naming
No built-in project management integration
Frequently asked questions
Everything you need to know about ChatML and how it might fit into your workflow.
Give it a try
We built ChatML for ourselves and hope you enjoy it too. Available on macOS, Windows, and Linux.
macOS 13.0+ · Windows 10+ · Linux x64