OpenRig vs Claude Managed Agents

Published April 2026 · Updated November 2026

OpenRig lets you work with a team of real Claude Code, Codex and Pi agents: talk to any one, follow its tasks and keep the team across restarts. This page compares that way of working with Claude Managed Agents, an API-driven managed service.

Two different questions about who operates it

Claude Managed Agents started from: what if running Claude agents as coordinated infrastructure could be a managed cloud service that ISVs and product teams integrate through an API? The answer is a managed agent harness. You configure agents with tools, MCP servers and skills, then drive them through API calls and event streams. Tools can run in Anthropic’s cloud or in a self-hosted sandbox; your application starts and directs the agents.

OpenRig started from a different question: what if I could run the coding agents I already use as one team, give them roles and talk to any of them? Define the team in YAML or start with a ready-made team. Each agent has its own conversation and address. You can work from the Claude Code or Codex session you're working in through the openrig plugin, see the whole team in the TUI, or open an agent's terminal.

Neither answer is obviously right. They're different bets about who the operator is. Anthropic bets on the application developer who wants Claude as a managed service inside their product. OpenRig is for developers who want a team on machines they run, with conversations they can join and tasks they can follow.

Side-by-side

Axis	OpenRig	Claude Managed Agents
What the primitive is	A rig: a team of coding agents with names, roles and their own conversations.	The agent definition: a versioned spec hosted in Anthropic's cloud, invoked through an API.
What runs as a node	Claude Code, Codex, Pi and Oh My Pi as they ship, plus community presets for other agents.	Claude agents with tools running in cloud or self-hosted sandboxes.
Where it runs	On your computer or another machine you choose; you run the team.	Anthropic manages the agent harness; tools can run in its cloud or a sandbox on your infrastructure.
Who operates it	You and your agents, through the openrig plugin, the TUI or agent terminals.	An application's code, calling the API and consuming SSE event streams.
How agents coordinate	rig send for questions; rig queue for tasks that stay assigned until they're done, handed from agent to agent.	An agent delegates within a session. Agents have separate conversation threads and share that session’s sandbox and filesystem.
License	Apache 2.0.	Proprietary cloud service.

Why I publish comparisons. This category is new. The vocabulary isn't settled. Nobody has a complete mental model of what's in the space, including me. The comparisons here are an attempt to find the shape of the category: what approaches are being taken, what each approach looks like, what we might learn from each other. Not a bake-off.

How you work with the team

Because OpenRig runs on your machine and you operate it directly, a few things fall out:

Start with a ready-made team. The operator can startWorkshop, a lead, builder, QA and reviewer, for ongoing work in your repo. Give them a goal, then change the team as the work grows.
Work from the agent you already use. Withthe openrig plugin, the Claude Code or Codex session you're working in can hand the team a task, follow it and read the result. To answer the team's questions from there, give the plugin a seat first. Each teammate also has a terminal you can open.
Choose an agent for each role. Claude Code, Codex, Pi and Oh My Pi can work together. Community presets support other agents, with the same team addresses and queue.
Bring the sessions you already have.rig discover finds coding-agent sessions you started in tmux; rig adopt gives them places on a team without starting their work again.
Come back to the same team.rig down saves a snapshot;rig up <name> --existing brings the team back and reports how each conversation returned. When the team needs your decision, the Feed and theSlack connector bring the question to you.

None of this is in opposition to Managed Agents. It's just a different shape for a different operator. If you're building a product and need Claude agents as a managed service you call from application code, Managed Agents is the right tool. If you want a team of coding agents you can talk to, give work to and keep across restarts, that's OpenRig. The category is large enough to have both shapes, and probably more.

Other Comparisons