| [ Web Proxy ] |
| Viewing: https://shopify.dev/docs/apps/build/ai-toolkit#example-migrate-extensions-with-the-toolkit | [Back] [Original] |
AI coding agents can search the web, read documentation, and inspect your codebase, but they still need the right context at the right time to be effective. Shopify development depends on platform-specific details: which API surface applies, what fields exist in a schema, which CLI command to run, and how an app or extension configuration should be validated.
Shopify AI Toolkit gives your AI coding tool a Shopify-aware starting point. The toolkit connects your agent to:
Before you install the Shopify AI Toolkit, make sure you have:
Plugins are the recommended way to install the AI Toolkit. The plugin bundles everything into a single install and updates automatically as new capabilities are released.
In your terminal, run claude plugin install:
In your terminal, run codex plugin add:
In your terminal, install the Shopify plugin:
In Cursor Chat, add the Shopify plugin:
In your terminal, download the install script and run it:
Ensure the Agent plugins preview is enabled in your VS Code settings.
Open the Command Palette (Cmd+Shift+P on macOS, Ctrl+Shift+P on Windows/Linux) and run:
When prompted, enter the repository URL:
In your terminal, run claude plugin install:
If you prefer to pick specific capabilities, you can manually add individual agent skills. You can browse the full list of available skills on GitHub. Note that manually added skills don't auto-update, so you'll need to pull updates yourself.
To install all skills:
To install a single skill, use the --skill flag. For example, to install only the GraphQL Admin API skill:
If you prefer MCP, you can connect to Shopify's developer resources through the Dev MCP server. The server runs locally and doesn't require authentication.
In your terminal, tell claude to add the MCP server:
Restart Claude Code to load the new configuration.
Add this configuration to your ~/.codex/config.toml file:
Codex uses TOML format with mcp_servers (snake_case) instead of JSON with mcpServers (camelCase). For more information, see the Codex MCP documentation.
Codex uses TOML format with mcp_servers (snake_case) instead of JSON with mcpServers (camelCase). For more information, see the Codex MCP documentation.
Restart Codex to load the new configuration.
Add this configuration to your ~/.gemini/config/mcp_config.json file:
Restart Antigravity CLI to load the new configuration.
Open Cursor and go to Cursor > Settings > Cursor Settings > Tools and MCP > New MCP server.
Add this configuration to your MCP servers (or use this link to add it automatically):
If you see connection errors on Windows, try this alternative configuration:
Save your configuration and restart Cursor.
Open VS Code and open the Command Palette (Cmd+Shift+P on MacOS, Ctrl+Shift+P on Windows/Linux).
Search for and select MCP: Open User Configuration.
Add this configuration to your user-level mcp.json file:
Save your configuration and restart VS Code.
In your terminal, tell claude to add the MCP server:
Restart Claude Code to load the new configuration.
One clear use case is extension migration. Checkout and customer account Shopify Extensions on API versions before 2025-10 need to adopt Polaris web components when upgrading to newer API versions.
These migrations involve repetitive, detail-heavy work:
shopify.extension.toml.With the toolkit installed, you can ask your agent to handle the migration. For example:
The toolkit doesn't remove the need to review and test the result. It helps automate the bulk of the migration work so you can spend more time validating behavior and preparing the updated extension for release.
After installing, your AI tool automatically draws on Shopify's developer resources when you ask it Shopify-related questions. Ask it to:
| Web Proxy Viewer | New URL | Original Page |