Questions and answers
The topic groups below answer common questions about using AgentGrid.io.
Getting started and plans
What is the fastest way to start?
Connect the AI you already use. Browser assistants use their official connector or plug-in; an agent with a shell can run npx @animaapp/cli@latest login. See How agents connect.
What does AgentGrid.io cost? You can start on Free without a card. Paid plans are Lite, Pro, and Business. Billing is per team and agents do not consume seats. See Plans and usage and the live pricing page.
How many URL or Figma imports are free? A Free team gets five successful imports total across both methods, not five of each. Deleting an import does not return it to the allowance. See Limits and quotas.
Teams, workspaces, and private access
What is the difference between a team and a workspace? A team contains members, workspaces, one billing plan, and one agent list. A workspace is a place inside the team that contains artifacts and has its own access list. See Workspace overview.
Can everyone on a team see every artifact? No. Visibility follows workspace access. A person or agent sees artifacts only in reachable workspaces.
Does an artifact link grant access? No. It identifies the artifact. The signed-in account or connected agent must also have workspace access. See Request access to a private artifact.
Where do I add people or agents?
Use team management for the team roster, which is the list of members. Use workspace access controls for workspace grants; each grant gives access to one workspace. Use agent administration for invitations, accountability, capabilities, and revocation. A capability allows an agent action such as read or write. Do not use the artifact Share panel as a team or agent roster. See Team management, Workspace access, and Invite an agent.
Can an artifact move between workspaces? Use the artifact move workflow when you have the required source and destination permissions. Moving between workspaces is different from duplicating between teams. See Rename, move, and duplicate.
Artifacts and editing
Do I have to save every Markdown edit? No. Markdown edits autosave continuously. The save indicator reports the current state. Save or Command-S creates an explicit history checkpoint; it is not required for routine persistence. See Edit a Markdown artifact.
Can several people edit at once? Yes. Live collaboration shows presence and remote selections where available. If a reconnect or conflict produces an unwanted saved result, use History to inspect and restore.
Can I undo a saved change? Yes. Open History, preview an earlier version, and restore it. The restore creates a new history entry.
Are folder and ZIP imports supported?
Yes. Import one folder or .zip, up to 50 MiB and 1,000 files. Include source and assets; omit dependencies and build output. See Import a folder or ZIP.
Why is an imported app blank?
It may be missing an entry point, such as index.html, or may have been imported as the wrong artifact type. Markdown imports need .md files. Correct the source and create a new import.
Are URL and Figma imports synchronized with their source? No. Each creates an editable AgentGrid artifact; later source changes are not pulled automatically.
Sharing and publishing
What is the difference between sharing and publishing? Sharing controls access to an artifact inside AgentGrid.io. Publishing creates a public URL. Publication is explicit and can expose content to anyone on the internet. See Share an artifact and Publish to a public URL.
Can Markdown be published? Yes. Publishing supports Markdown as documented on the publish page.
How do I update or unpublish a site?
Publish the current artifact state to update it, or unpublish to take the public site offline without deleting the artifact. Agents need publish and must act only after an explicit human request.
Agents and CLI
Does connecting an agent give it my password or consume a seat? No. It gets a separate identity, never receives your password, and does not consume a seat.
What determines what an agent can see? Its workspace grants, capabilities, and the current permissions of the person or team accountable for it. A personal agent is accountable to one human. A team agent is accountable to the team. Workspace access is separate from this accountability. See Agent identity and stewardship.
How does an agent find workspaces?
Run npx @animaapp/cli@latest workspaces. Use the returned workspace ID with list --workspace, create --workspace, or duplicate --workspace. See The Anima CLI.
What happens if no workspace is supplied when creating? It works only when the agent can create in exactly one workspace. With several writable workspaces there is no default, so the ID is required.
Why did an agent invite connect nothing? Creating an invite is only advance consent. Copy and deliver the one-time link; the agent must redeem it.
Can an agent publish on its own?
It may publish only when it has publish and a human explicitly asks. A capable agent should offer publishing as a follow-up, not assume consent.
Claiming and support
Why did my AI send a claim link? It created anonymously before connecting. Claim within 24 hours of creation or the content is removed. See What is claiming.
How do I contact a person?
Email [email protected]. Include the artifact, team, workspace, time, and exact error, but never credentials. See Contact support.