Collaborate with your agents.
And everyone else's.
A rig is a shared workspace your agents already know how to use. Everyone you invite works in it together, live. Every version, every decision, and the reason behind it stays with the work.
Building is the easy part now. Keeping every person and agent on the same context isn't.
Context gets lost
Workflows in one app. Data in another. Prompts in a third. Nothing connects. You re-explain yourself every session.
Agents drift without structure
No memory between sessions, no consistent workflows, no way to encode what "good" looks like.
Trapped on one machine, one agent
Your setup is only as good as the laptop it’s on. Switch machines, lose context. Switch agents, redo everything from scratch.
More output than you can check
Every agent ships 100x more, and all of it lands on you to review. The bottleneck moves from building the work to holding the bar on it.
Sharing work is painful
You built something great. Now try giving it to a non-technical colleague. Good luck with the README and the setup script.
Your team can’t work together
Different copies of the same agent on every laptop. Edits don’t propagate. Teammates work in parallel-but-disconnected. Shared-doc apps don’t apply to agent folders.
Introducing rig.
One folder you own, run anywhere, and share live with your team.
Files and folders, not code
Lay out what your agent knows, how it works, and what it remembers, as files in a folder. Plain text. Full control.
Works with your setup
Use Claude Code, Codex, or any agent you already have. Rig doesn’t replace your tools. It gives them structure.
Share your rig live
Connect your rigs into one living structure that mirrors how you work, by project, team, or org. Context and work flow across every person and agent live, so coordination happens on its own.
Use what others built
Browse the Hub. Find a rig for your use case. Install it with one command. Skip the setup, start the work.
Four steps to a rig everyone works in. From your folder to your whole team.
Everything your agents share.
Built in.
Comment on the exact line
Open a thread on a passage and bring any agent into it. Run several at once, each on the precise context it needs, and the conversation stays where the work is.
The why behind the work
What you see is the latest version. Underneath sit every earlier version and the session behind each change: the prompt, the model, the intent. Nobody has to guess why something is there.
One link, and they’re in
Share a rig with a person or an agent. Either one opens the link and starts working. No walled garden, nothing to set up first.
Skills travel with the rig
How you answer matters as much as what you answer. The skills and shared context a team works from are attached to the rig, and every agent that opens it uses them.
Auto-configured environment
APIs, databases, and services, connected out of the box. Rig manages MCP connections and auto-configures your agent’s environment.
Share the structure, keep the secrets
Structure over content when sharing. Share the safe files; keep credentials and private data local-only. You decide what travels and what stays put.
Rig bends to your structure. Your agents work the way you operate, not the other way around.
Set up a rig per project and one for the team. The team's agents see what each project's people and agents are doing, as it happens; a project's agents draw on the team's standards. Structure your rigs to mirror how your org actually works: nested, distributed, and fluid.
Find the right agent.
Skip the setup.
Browse rigs published by the community. Install any of them with a single command and start working immediately.