Orchestration vs coordination
Do you need an orchestrator for multiple coding agents?
Orchestrators decide what runs. Coordination keeps agents that are already running from conflicting. Most teams with several coding agents need the second first.
The problem
Teams reach for a framework that spawns and directs agents, when the real pain is that the agents they already run can't see each other.
With Concord
01
Name the pain
If agents collide, duplicate work or lose decisions, that's coordination. If you need agents started automatically in a fixed pipeline, that's orchestration.
02
Keep your harnesses
A coordination layer works with Claude Code, Codex and Cursor as they are, so nobody has to change how they drive their agent.
03
Add structure where needed
Claims, messages, assignments and handoffs give the structure people want from an orchestrator, without a central process owning every agent.
Result
Agents stay independent and still behave like a team.
Common questions
- Is Concord an orchestrator?
- No. It doesn't start, schedule or direct agents. It is the shared state and messaging they use while you, or an orchestrator, run them.
- Can I use both?
- Yes. An orchestrator that launches MCP-capable agents can have them claim and record work through Concord like any other agent.
Further reading
Give your agents shared work-state.
npm install -g @concord-ai/concord-mcp