Skip to main content
Moda for Agents lets the agent you already use make real, editable Moda designs — decks, one-pagers, social posts, diagrams, websites, video — and hand you back a canvas link plus the exported file. Not a screenshot, not a flat image: every layer stays live and editable at moda.app. There are two ways in, and picking is easy:
  • Your agent can run commands (Claude Code, Codex, Cursor, and friends) → install the moda CLI. This is the better experience: cheaper in context, faster, versioned, and it can read the files on your machine.
  • Your agent lives in a browser (claude.ai, Claude Desktop, ChatGPT) → add the Moda connector. Same account, same designs, nothing to install.
Both talk to the same Moda service, so entitlements, credits, brand kits, and canvases are identical either way. There is no transport tax.

​
Pick your setup

​
What your agent can make

​
How the pieces fit

Connect

Install the CLI or add the connector. One sign-in either way.

Teach

Install the Moda skills so your agent knows Moda’s design workflow instead of guessing at it.

Ask

Describe the artifact you want. Stuck mid-task, your agent can call ask_expert — free, grounded answers about how to do something in Moda.

Ship

You get a canvas URL you can keep editing, plus the exported file.

​
Going further

  • CLI quickstart — install, sign in, first canvas.
  • MCP connector — the claude.ai and Claude Desktop setup.
  • Skills — how your agent learns Moda’s workflow.
  • ask_expert — the built-in advice lane.
  • API reference — the HTTPS surface both the CLI and the connector are built on. Use it directly when you are writing the integration yourself.
Looking for the older mcp.moda.app MCP server (start_design_task, get_moda_canvas)? Those docs are still published under MCP Server (Legacy). New integrations belong here.