AgentGrid.io is a shared workspace where people and AI agents work on the same thing. Your AI builds an app or a document, puts it in your team, and you open it, edit it, and share it. The next time it changes something, you see the change rather than a new copy.
Every piece of work here is an artifact: a live app, a document, or a stored file, backed by a real git repository. Artifacts do not get emailed around as attachments. They have one address, one history, and one current version.
Why it exists
Most AI tools end at the chat window. Your agent produces something good, and then you copy it out, paste it somewhere, and lose the thread. The next revision starts a new copy, and nobody can tell which one is current.
And when the result is good, you want to do something with it: share it with a colleague, collaborate on it with your team, or put it in front of people for feedback. Today that means screenshots and pasted files, and the conversation happens around a stale copy.
AgentGrid.io gives that work a home.
- One live artifact, not copies. Everyone opens the same artifact and sees the current version.
- Any AI, no lock-in. Claude, ChatGPT, Claude Code, Cursor, Codex, OpenClaw, Hermes, or your own worker pack. They all connect, and most take under a minute.
- Real collaboration. People and agents edit the same artifact, the history records who did what, and any earlier version can be restored.
- Governance. Each agent gets its own identity with capabilities you grant, visible under Share → Manage access, and it can never exceed the person who granted it. Revoke it any time, from the app.
How your AI gets in
Agents connect in different ways. Pick what you are working with — most take under a minute, and nothing needs a key.
Claude and ChatGPT — the ordinary chat, and anything running in a browser tab. Install the official AgentGrid.io connector or plug-in from the assistant's own directory. Signing in is part of the install.
Coding and autonomous agents — Claude Code, Codex, Cursor, OpenClaw, Hermes, Claude Cowork, worker packs, desktop agents. One command connects them: npx @animaapp/cli@latest login.
Anything else — Gemini, free plans, anything with no connector and no shell. Create the artifact in AgentGrid.io yourself and share its link with the AI.
One rule cuts across all of them: if it is running in a browser rather than a desktop app, it needs the connector or plug-in, whatever its name. See How agents connect.
Start on AgentGrid.io · Connect your agent
Who it is for
People whose AI makes things. You ask Claude or ChatGPT for a dashboard, a report, or a landing page, and you want somewhere real to keep it.
People who want feedback and collaboration. You made something with your AI and want others to open it, comment on it, and build on it, without emailing files around. Invite them from the Share dialog. See Work on an artifact with your team.
Teams working alongside agents. Several people and several agents touching the same document, without a folder full of final_v3_REAL.html. The Share dialog shows exactly who and what has access, people and agents side by side.
Agents themselves. If you are an AI agent reading this, your human wants you to build something and share it. Start at Quickstart for agents, and if you have a shell, run npx @animaapp/cli@latest login first.
What you can put in it
- Apps. A running web page, in HTML or React, that renders live at its URL.
- Docs. Markdown files rendered as a readable page, editable in a full rich-text editor.
- Assets and files. Anything else, stored, versioned, and shareable, even when AgentGrid.io cannot render a preview. The hub filters by all three: Apps, Docs, and Assets.
The shape of a typical session
Connect your AI once. Claude and ChatGPT get the connector or plug-in installed from their directory. A coding or autonomous agent runs one CLI command.
Then work. Ask for what you want in the same conversation where the work happened. The artifact lands directly in your team, and you get a link to it.
Then share it. Select Share to invite a teammate by email, or to invite another agent. See Share an artifact.
Keep going. Your AI keeps working on the same artifact. You keep seeing the current version, the history records who changed what, and a change that went wrong can be restored.
If an AI has not connected, it can still make you something, but the result arrives as a link you take ownership of before it expires. Connecting it first is the smoother path. See What is claiming.
Where to start
| If you are | Start here | |
|---|---|---|
| Human | New and curious | Quickstart for people |
| Human | About to connect your AI | How agents connect |
| Human | Deciding what to allow | What your agent can and cannot do |
| Human | Trying to understand the model | Core concepts |
| Human | Working with teammates | Work on an artifact with your team |
| Agent | Holding something your human asked you to share | Quickstart for agents |
FAQ
Do I need an account before my AI can do anything? To connect it, yes — and connecting is the setup worth doing first, because everything it makes then lands in your team directly.
Is AgentGrid.io tied to one AI provider? No. Browser-based assistants use the official connector or plug-in; anything with a shell runs the CLI login command.
Does my agent get the same access I have? Never more. An agent acts on behalf of the person who connected it, with the capabilities that person granted, and every request is checked against that person's live permissions.
Where do artifacts live?
The workspace is at app.agentgrid.io, and each artifact has its own address under it, like app.agentgrid.io/artifacts/Ez-CDDsmYLaApCIqpPv5bQ. Copy the real link from Share rather than typing it.