Gastown Alternative — AI Agent Orchestration That Ships

Why Developers Switch

What chAIrman does differently

Most AI coding platforms give you a single agent in a single context window. That works for small edits. It breaks when you need a database engineer, an API developer, a frontend builder, and a QA engineer all working on the same feature.

chAIrman is an MCP server that turns Claude Desktop into a CEO managing a full engineering team. You get structured backlogs, agent pipelines, file ownership enforcement, auto-retry with handoff inheritance, a skills library, alumni archives, and per-project budget controls.

Unlimited concurrent agents with smart scheduling based on CPU and memory
54 MCP tools for project management, team management, messaging, metrics, deployment, and more
10 pre-built agent templates: frontend-lead, backend-lead, qa-engineer, devops, security-auditor, and more

// chAIrman: manage an AI team // Hire from pre-built templates hire_from_template({ template: "backend-lead", project: "my-saas", tech_stack: "Node.js, Express, PostgreSQL" }) hire_from_template({ template: "frontend-lead", project: "my-saas", tech_stack: "React, TypeScript, Tailwind" }) hire_from_template({ template: "qa-engineer", project: "my-saas", tech_stack: "Playwright, Jest" }) // 3 agents with full job descriptions // auto-populated from templates

Head-to-Head

chAIrman vs. typical AI coding platforms

Side-by-side comparison of what you get with chAIrman versus single-agent AI coding tools.

Capability	Single-Agent Tools	chAIrman
Concurrent agents	1	Unlimited (budget-gated)
Dependency management	Manual	Automatic pipelines + cycle detection
Failure recovery	Start over	Auto-retry with handoff inheritance
Context injection	Copy-paste	CLAUDE.md + skills + messages auto-injected
Cost tracking	API dashboard	Per-agent, per-task, per-project with budget limits
Monitoring	Terminal output	Web dashboard + WebSocket + desktop notifications
Knowledge preservation	Lost between sessions	Alumni archive + veteran rehiring
Git integration	Manual commits	Auto-commit on task completion

Built for Teams

Features that matter when shipping

Everything you need to manage AI agents like a real engineering team.

1

Structured Backlogs

Every project starts with a backlog of milestones and tickets. Each ticket defines what to build, which files to touch, what's off-limits, and how to verify success. You plan before agents write code.

2

Live Dashboard

A web dashboard at localhost:3456 with kanban board, live terminal view, pipeline visualization, budget tracker, message board, git history, and alumni browser. Real-time updates via WebSocket.

3

857-Skill Library

24 skill categories covering frontend design, backend architecture, testing, DevOps, security, databases, APIs, documentation, and more. Skills auto-inject into agent prompts via TF-IDF matching.

4

Webhook Integration

Register URLs to receive event notifications: agent hired, task completed, task failed, budget warning, pipeline complete. Optional HMAC-SHA256 signing for request verification. Connect to CI/CD pipelines.

5

Health Monitoring

System-level health checks for memory, CPU, zombie detection. Project-level checks for error rates, cost velocity, blocked pipelines. Returns HEALTHY, WARNING, or CRITICAL with actionable details.

6

Export & Import

Export a project's full state (agents, handoffs, alumni, messages, notes, timeline, metrics) as JSON. Import it on another machine or share with a teammate. Optional gzip compression.

54

MCP Tools

10

Agent Templates

857

Skills Injected

7

Webhook Events

Common Questions

Switching to chAIrman FAQ

How does chAIrman compare to Gastown?

chAIrman is an MCP server that turns Claude Desktop into a CEO managing a full AI engineering team. You get structured backlogs, dependency pipelines, auto-retry with handoff inheritance, an 857-skill library, alumni archives, budget enforcement, webhooks, health monitoring, and a real-time dashboard. It's designed for managing multiple agents on real features, not just single-agent code edits.

What models does chAIrman support?

chAIrman supports all Claude models available through Claude Code. Use model aliases: "sonnet" for fast, cost-effective tasks; "opus" for complex architecture, refactoring, and security work. You can set a default model per agent and override it per task. The smart scheduler can also recommend the optimal model based on task keywords.

How much does it cost?

chAIrman plans start at $19.99/month for Pro (2 concurrent agents, 5 projects) and $99.99/month for Unlimited (unlimited agents and projects). AI compute costs from your Anthropic API usage are separate. chAIrman tracks all costs per-agent, per-task, and per-project so you always know what you're spending.

Can I migrate existing projects?

Yes. Create a project pointing to your existing directory, and agents can read and modify your codebase immediately. If you have CLAUDE.md files, they're auto-injected into every agent's prompt. The export/import system also lets you move projects between machines.

How do I get started?

Sign up at mrchairman.ai/account, download the binary for your platform, and add chAIrman as an MCP server in your Claude Desktop config. The setup guide at mrchairman.ai/setup walks through every step. You'll be hiring agents in under 5 minutes.

Is my code sent to external servers?

Agents run as local Claude Code processes on your machine. Your code stays on your machine. chAIrman's MCP server runs locally. The only external calls are to the Anthropic API (for Claude) and to api.mrchairman.ai for prompt assembly and license validation. No source code is sent to chAIrman's servers.

The Gastown Alternative That Actually Ships