We had 20 Claude terminals open. So we built an orchestration layer.
Source: Dev.to
Background
Last month, my co‑founder Ciprian and I hit peak chaos.
We had 20 Claude Code terminals running—each working on a different part of our codebase. Tab after tab, terminal after terminal.
“Wait, which agent was working on the API?”
We were spending more time managing AI agents than actually building. The irony wasn’t lost on us.
Our Workflow
- Open Claude terminal #1 for frontend
- Open Claude terminal #2 for backend, and so on…
We thought: if AI agents are supposed to 10× our productivity, why does managing them feel like a full‑time job?
The Orchestration Layer
We built what we needed—a single dashboard that coordinates multiple agents, eliminating the chaos.
How It Works
-
Specialist agents (e.g., Database Expert, React Developer, API Developer) each have their own terminal, branch, and expertise.
-
Define task flows between agents with a drag‑and‑drop interface:
Frontend Dev → React Developer → API Developer -
When one finishes, the next picks up automatically.
-
Security: API keys stay on your machine; your code never touches our servers (100 % local, 100 % private).
Key Learnings
- Specialists outperform generalists – a “React Developer” agent writes better React code than a generic agent.
- Parallel > Sequential – five agents working simultaneously beat one agent handling five tasks.
- Visibility matters – you can’t manage what you can’t see; a unified dashboard changes everything.
Private Beta
We’re currently in private beta, and it’s free.
👉 orcha.nl
We’d love feedback from the developer community. Drop a comment or reach out—let’s figure this out together.
Built by David & Ciprian, two devs who got tired of tab chaos.