ShotCanvas STUDIO Open Studio →
Guide · Model Context Protocol · Updated June 2026

App Store screenshots, from your AI assistant.

ShotCanvas ships a remote MCP (Model Context Protocol) server, so Claude, ChatGPT, and any MCP-compatible client can design store screenshots, write the metadata, grade your set, and publish to the App Store and Google Play — all from a normal chat. Just say: “design App Store screenshots for my app and write the listing.”

What it is

The Model Context Protocol is the open standard AI assistants use to connect to external tools. ShotCanvas exposes its whole screenshot-and-ASO toolkit over a hosted MCP endpoint, so an AI can do the work for you instead of describing how to do it. It's a remote server (Streamable HTTP) — nothing to install — that connects to Claude (Code, Desktop, API), ChatGPT and the OpenAI Responses API, and any other MCP client.

What the AI can do

design_screenshotsAI-designed App Store / Play Store screenshots from your raw captures.
render_screenshotRender one screenshot in a specific template and size.
grade_screenshotsAn honest letter grade for your set, with ranked fixes.
generate_metadataTitles, subtitles, keywords, descriptions — within store limits.
suggest_keywords / headlinesASO keyword research and conversion headlines.
localize_metadataTranslate your listing into any market.
publish_to_app_storeUpload straight to App Store Connect.
publish_to_google_playUpload straight to Google Play.

Get your connection

Open the ShotCanvas Studio, sign in, and choose Setup MCP from the account menu. You'll get your personal server URL and access token, with copy-paste setup for every client. Use them in the snippets below in place of <YOUR_URL> and <YOUR_TOKEN>.

Connect it

Claude Code (CLI)

claude mcp add --transport http shotcanvas <YOUR_URL> \
  --header "Authorization: Bearer <YOUR_TOKEN>"

Claude Desktop

Add to claude_desktop_config.json (uses the mcp-remote bridge), then restart:

{
  "mcpServers": {
    "shotcanvas": {
      "command": "npx",
      "args": ["mcp-remote", "<YOUR_URL>", "--header", "Authorization: Bearer <YOUR_TOKEN>"]
    }
  }
}

Claude API (Messages connector)

"mcp_servers": [{
  "type": "url",
  "url": "<YOUR_URL>",
  "name": "shotcanvas",
  "authorization_token": "<YOUR_TOKEN>"
}],
"tools": [{ "type": "mcp_toolset", "mcp_server_name": "shotcanvas" }]

ChatGPT (Developer mode)

Settings → Apps & Connectors → enable Developer mode → Add custom connector → paste <YOUR_URL> and your token. Then ask ChatGPT to design your screenshots.

OpenAI Responses API

"tools": [{
  "type": "mcp",
  "server_label": "shotcanvas",
  "server_url": "<YOUR_URL>",
  "authorization": "<YOUR_TOKEN>",
  "require_approval": "never"
}]
Any MCP client works. ShotCanvas speaks Streamable HTTP — point any Model-Context-Protocol-compatible tool at your server URL with the bearer token and the full toolkit appears.

Try saying

Prefer a visual editor? Everything the MCP can do is also in the browser Studio and the ShotCanvas mobile app. New to store images? Start with screenshot sizes or designs that convert.

Open Studio → Setup MCP Compare AI screenshot tools