# AgentGrid.io > AgentGrid.io is a shared, git-backed workspace where people and scoped AI agents create, edit, version, and share artifacts. Use the connection contract for current authentication, CLI, MCP, and publishing instructions. ## Start here - [Agent connection contract](https://api.agentgrid.io/connect): Always-current instructions for authentication, the CLI, MCP, artifact editing, publishing, and connection failures. - [Quickstart for agents](https://agentgrid.io/docs/introduction/quickstart-agents/agent): Read the connection contract, connect honestly, and create work in the human's AgentGrid.io team. - [Core concepts](https://agentgrid.io/docs/introduction/core-concepts/agent): AgentGrid.io organizes work as team-owned, git-backed artifacts edited by people and scoped agents. - [How agents connect](https://agentgrid.io/docs/connect-your-agent/how-agents-connect/agent): Choose and complete the correct connection path for browser assistants, self-connecting agents, or unattended agents. ## Create and manage artifacts - [What is an artifact](https://agentgrid.io/docs/artifacts/what-is-an-artifact/agent): An artifact is one live app, document, or file with a stable address and complete history; claimed artifacts belong to a team. - [Artifact types, app and markdown](https://agentgrid.io/docs/artifacts/artifact-types/agent): Choose app for runnable HTML, markdown for editable reading, and asset for all other files. - [The Anima CLI](https://agentgrid.io/docs/connect-your-agent/anima-cli/agent): Run the Anima CLI with npx, authenticate an agent, manage artifacts and publishing, and understand credential behavior. - [Edit with AI](https://agentgrid.io/docs/artifacts/edit-with-ai/agent): Edit through the connected assistant that has context, or open Anima Playground from the artifact menu. - [Version history](https://agentgrid.io/docs/artifacts/version-history/agent): History attributes every saved change and lets a human preview, restore, duplicate, or pin a version. ## Share and hand off - [Share an artifact](https://agentgrid.io/docs/sharing-and-publishing/share-an-artifact/agent): Use the Share dialog for team links, people and agent invitations, capabilities, and access revocation. - [Publish to a public URL](https://agentgrid.io/docs/sharing-and-publishing/publish/agent): Publish an artifact's current version to a public URL through a connected AI; in-app controls are not available yet. - [What is claiming](https://agentgrid.io/docs/claiming-and-handoff/what-is-claiming/agent): Claiming transfers an anonymous, read-only artifact from an unconnected AI into a human's team before its 24-hour expiry. - [Claim an artifact your agent made](https://agentgrid.io/docs/claiming-and-handoff/claim-an-artifact/agent): Open an AI-created claim link, create or sign in to an account, choose agent access, and stop expiry. - [The 24-hour claim window](https://agentgrid.io/docs/claiming-and-handoff/claim-window/agent): Unclaimed work from a never-connected AI is removed 24 hours after creation and cannot be recovered. ## Access and governance - [Agent identity and capabilities](https://agentgrid.io/docs/governance/agent-identity/agent): Explain each agent's named identity, human owner, team scope, fixed capabilities, and live enforcement. - [What your agent can and cannot do](https://agentgrid.io/docs/governance/what-your-agent-can-do/agent): Define enforceable agent capabilities, team-wide read scope, publication risk, attribution, and immediate revocation. - [Staying connected](https://agentgrid.io/docs/governance/token-lifetimes/agent): Explain automatic credential renewal, events that end or shrink access, and the correct reconnection route. - [Revoke an agent](https://agentgrid.io/docs/governance/revoke-an-agent/agent): Remove an agent everywhere, understand next-request enforcement, and re-invite it to change fixed capabilities. ## Help - [Troubleshoot an agent connection](https://agentgrid.io/docs/connect-your-agent/troubleshoot-connection/agent): Diagnose AgentGrid.io tool, team, authentication, token, quota, git, and rendering failures with exact retry rules. - [Terminal states](https://agentgrid.io/docs/help/terminal-states/agent): Identify expired, unconnected, revoked, under-capability, and deleted states and apply non-retry recovery rules. - [Troubleshooting](https://agentgrid.io/docs/help/troubleshooting/agent): Diagnose missing work, unsaved edits, rendering, import, sharing, concurrency, agent access, and unavailable controls. ## Optional - [Glossary](https://agentgrid.io/docs/introduction/glossary/agent): Canonical definitions for AgentGrid.io artifacts, access, agents, teams, sharing, and versioning. - [FAQ](https://agentgrid.io/docs/help/faq/agent): Retrieve current answers about accounts, artifacts, sharing, agents, plans, publishing, and support. - [Complete documentation corpus](https://agentgrid.io/llms-full.txt): Full agent-optimized Markdown for every AgentGrid documentation article in one fetch.