Find the right component
Start with the screen you need. The agent will search the catalog before coding.
Use the avada-polaris MCP to find the best components for a Shopify app settings page. Explain your choices before you write code.
The gallery serves itself to AI coding agents over MCP, so an agent can read the real component sources instead of writing Polaris from memory — and publish new components back. Sign in to get your own token.
GOOGLE_CLIENT_ID, GOOGLE_CLIENT_SECRET, AUTH_ALLOWED_DOMAIN and AUTH_SESSION_SECRET, or ask an admin to issue a token with the create_token tool.Cursor (.cursor/mcp.json) and Claude Desktop (claude_desktop_config.json):
{
"mcpServers": {
"avada-polaris": {
"type": "http",
"url": "<this site>/api/mcp",
"headers": { "Authorization": "Bearer pcm_..." }
}
}
}How to use MCP
Once connected, mention avada-polaris MCP in your request. Describe the screen and constraints; your coding agent can search the catalog, read the exact source, and compose the result.
Say what you are building and who will use it.
Include states, responsiveness, and your desired output.
Tell the agent to search and reuse the real component sources.
Copy one, then replace the details with your own.
Start with the screen you need. The agent will search the catalog before coding.
Use the avada-polaris MCP to find the best components for a Shopify app settings page. Explain your choices before you write code.
Name the page, the content it needs, and the output you expect.
Using avada-polaris MCP, build a responsive settings screen with an onboarding checklist, a rules table, and pricing plans. Use the real component sources and return a runnable React implementation.
Give the agent context about the product and ask it to reuse catalog patterns.
I am building a Shopify app onboarding flow. Search avada-polaris for reusable patterns, then compose a page with setup progress, clear next actions, loading and empty states. Follow the component APIs exactly.