AI & MCP guides

MCP for local runtime; ACP for Rig Cloud in the editor.

25 guides in this category

• Connect AI editors (MCP): Hub setup, tools list, and staying glued to the right project port.
• Connect editors via ACP: Agent Client Protocol for Rig Cloud — status, devices, Mesh, and tickets from Zed or JetBrains.
• What MCP is in Rig: Model Context Protocol: scoped tools on your real processes.
• Connect Cursor: One-click MCP link for Cursor.
• Connect Claude: Wire Claude to Rig’s MCP hub.
• Connect Codex: OpenAI Codex via Rig MCP.
• Connect VS Code: MCP for VS Code / Copilot-style workflows.
• Connect Windsurf: Link Windsurf to the Rig hub.
• Connect GitHub Copilot CLI: One-click link to Rig MCP for Copilot CLI.
• Connect OpenCode: Link OpenCode to Rig's local MCP hub.
• Connect Pi: Set up Pi with Rig MCP (streamable HTTP).
• Connect OpenClaw: Use Rig as a generic remote MCP server in OpenClaw.
• Connect Antigravity: Use Rig as a generic remote MCP server in Antigravity.
• Connect LM Studio: Attach LM Studio to Rig's MCP hub for tool calls.
• Connect Hermes: Use Rig as a generic remote MCP server in Hermes.
• MCP tools overview: What agents can call today.
• Project-scoped MCP: Keep the agent on the right repo and port.
• MCP open_url gotchas: Why Rig refuses the wrong localhost.
• MCP activity log: See what the agent called.
• Guidelines for agents using Rig: Prompt patterns that keep MCP reliable.
• What ACP is in Rig: Agent Client Protocol: editor ↔ agent — not the same as MCP.
• Connect Zed via ACP: Walkthrough: register Rig Cloud as an ACP agent in Zed settings.
• ACP setup (stdio and cloud): Local stdio from the repo, or the cloud service at acp.userig.app.
• ACP auth and tokens: Service bearer vs personal developer token — what each one unlocks.
• ACP vs MCP in Rig: When to use the local MCP hub vs the cloud ACP agent.

See all Rig documentation in the knowledge base.

https://userig.app/docs/c/ai