Mission Control — Self-Hosted AI Agent Orchestration Platform

Introduction

Mission Control gives teams a single dashboard for managing AI agent workloads. You define tasks, assign them to agents, set budgets and constraints, then monitor execution and costs in real time. It is designed for teams that run multiple agents across projects and need visibility into what each agent is doing and spending.

What Mission Control Does

• Provides a web dashboard for creating, dispatching, and monitoring AI agent tasks
• Supports multi-agent workflows where tasks are chained or run in parallel
• Tracks token usage and estimated costs per task, agent, and project
• Enforces budgets and governance policies before agents execute
• Stores execution history and outputs in a local SQLite database

Architecture Overview

Mission Control is a Next.js application with a TypeScript backend. The orchestration layer accepts task definitions via the UI or API, validates them against governance rules, and dispatches them to configured agent backends. Agent output is captured, analyzed for token usage, and stored locally. The dashboard renders real-time status via server-sent events and provides historical analytics through chart components.

Self-Hosting & Configuration

• Clone the repo and run with Node.js 18+ or deploy via Docker
• Configure agent backends (Claude, OpenAI, local models) in .env
• Set per-project budgets and token limits through the dashboard settings
• Data persists in SQLite by default; no external database required
• Deploy behind a reverse proxy with authentication for team access

Key Features

• Centralized dashboard for all agent activity across projects
• Built-in cost tracking and budget enforcement
• Multi-agent workflow support with task dependencies
• Governance rules that gate execution on approval or policy
• Self-hosted with no cloud dependency; all data stays local

Comparison with Similar Tools

• Paseo — mobile-first task dispatch; Mission Control focuses on team dashboards and governance
• Archon — harness building and workflow design; Mission Control emphasizes runtime orchestration
• LangSmith — cloud-hosted LLM monitoring; Mission Control is self-hosted with agent dispatch
• n8n — general workflow automation; Mission Control is purpose-built for AI agent orchestration

FAQ

Q: Does Mission Control run the agents itself? A: It dispatches tasks to configured agent backends (Claude Code, Codex, etc.) and monitors their execution.

Q: Can I set spending limits per project? A: Yes. Configure per-project token budgets and cost ceilings in the dashboard settings.

Q: Does it require a cloud account? A: No. Mission Control is fully self-hosted. You only need API keys for the LLM providers your agents use.

Q: Is it suitable for solo developers? A: Yes, though it provides the most value for teams managing multiple agents and projects.

Sources

• https://github.com/builderz-labs/mission-control