Agent Grid MCP Server Guide

Agent Grid is a governed hub, by Anima, where AI agents build, host, publish, and share web apps. This repo is the MCP server guide and the skill that teaches your coding agent to use it — install it as a Claude Code plugin, a Codex skill, or a plain MCP server.

Agent Grid lives at **app.agentgrid.io**. For the full docs, see our help documentation.

What your agent can do

• Build an app from a prompt, a website, or a Figma design — describe it, clone a URL, or hand over Figma frames, and get back a running app at a live URL.
• Host code you already have — import a project or start an empty repo, and push to it.
• Edit any artifact over git — every artifact is a real git repository; clone, commit, push.
• Publish to a live public URL — and take it down again.
• Generate design-aware code into your own codebase — Figma frames to files in your repo, matched to your stack.

Everything is an artifact: one git repo in your team's workspace, addressed by its `sessionId`.

Quick start: Claude Code plugin

/plugin marketplace add AnimaApp/mcp-server-guide
/plugin install anima@mcp-server-guide

This configures the MCP server and installs the skill. Authenticate when prompted. That's it.

Installation & setup

Requirements

• An MCP-capable AI coding tool — Claude Code, Codex, Cursor, VS Code, or anything speaking streamable HTTP
• An Agent Grid account

Server URL: `https://api.agentgrid.io/v1/mcp` · Transport: streamable HTTP

Claude Code (manual)

1. In your terminal (not inside Claude Code):

claude mcp add --transport http anima https://api.agentgrid.io/v1/mcp

2. Restart Claude Code

3. Run `/mcp`, select anima, and authenticate in the browser

4. (Optional) Connect Figma during authentication to enable the Figma flows

Manage servers with `claude mcp list`, `claude mcp get anima`, `claude mcp remove anima`. See Anthropic's MCP documentation.

OpenAI Codex

codex mcp add anima --url https://api.agentgrid.io/v1/mcp

Or in `~/.codex/config.toml`:

[mcp_servers.anima]
url = "https://api.agentgrid.io/v1/mcp"

Then install the skill (recommended):

codex skill install AnimaApp/mcp-server-guide/skills/anima

See OpenAI's Codex MCP documentation.

VS Code

1. `Cmd Shift P` / `Ctrl Shift P` → `MCP: Add Server`

2. Select HTTP

3. URL: `https://api.agentgrid.io/v1/mcp`

4. Server ID: `anima`

5. Choose global or workspace scope

{
  "servers": {
    "anima": {
      "type": "http",
      "url": "https://api.agentgrid.io/v1/mcp"
    }
  }
}

> [!NOTE]

> MCP in VS Code requires GitHub Copilot. See VS Code's documentation.

Cursor

Cursor > Settings > Cursor Settings > MCP > + Add new global MCP server:

{
  "mcpServers": {
    "anima": {
      "url": "https://api.agentgrid.io/v1/mcp"
    }
  }
}

Click Connect to authenticate. See Cursor's documentation.

Any other client

{
  "mcpServers": {
    "anima": {
      "url": "https://api.agentgrid.io/v1/mcp"
    }
  }
}

No MCP? Use the CLI

The Anima CLI runs the same tools from a shell — no MCP server to configure:

npx @animaapp/cli@latest login
npx @animaapp/cli@latest create -t p2c -p "SaaS dashboard with sidebar and analytics"

Prompting your agent

Build something new

> "Build me a SaaS analytics dashboard with a sidebar, KPI cards, and an activity feed."

Your agent creates an artifact, waits for the build, and hands back a live preview URL. Ask for variants and it can generate several in parallel to compare.

Clone a site or implement a Figma design as a live app

> "Clone stripe.com/payments into an app."

> "Turn this Figma frame into a working site: https://figma.com/design/..."

Edit an existing artifact

> "The header on https://app.agentgrid.io/artifacts/xyz is broken on mobile — fix it."

The agent clones the artifact's git repo, fixes it, and pushes. Content changes always go through git.

Implement a Figma design in your own codebase

> "Implement this Figma design in my project: https://figma.com/design/..."

The agent generates code matched to your stack and uses the returned design snapshots as visual reference.

Publish

> "Publish it."

Publishing makes an app public to the world — it isn't needed just to share, since the artifact URL is already viewable.

Design system access (Enterprise)

> "Implement a login form following our design system, using this Figma URL: ..."

Design system setup is done with our team. Contact us to get it configured.

Best practices

Describe intent, not CSS. Agent Grid is design-aware; hex values and pixel dimensions in a prompt override that and produce generic results. Say what it's for, who it's for, and the 3–5 features that matter.

Be specific when implementing. "Implement the login form using our existing Button and Input components from `src/components/ui`" beats "implement this design".

Break down large designs. Header first, then the card grid, then the footer. Focused requests are more accurate.

Troubleshooting

Common issues — async generation, create-type validation, git token expiry, publish semantics — are covered in the skill's troubleshooting reference.

MCP not recognized: confirm the server is configured in your tool and that authentication completed.

Authentication issues: remove and re-add the server, clear cookies, restart your client.

What's in this repo

Additional resources

• Agent Grid
• Anima MCP documentation
• Anima CLI
• Contact us for Enterprise features