Expokit

AI assistants (MCP)

The ExpoKit MCP server lets your AI editor look up components, read their props and examples, and add them with the CLI — instead of guessing an API or writing a component from scratch.

It runs on your machine through npx, reads the registry bundled in the CLI and never touches the network or your Pro token. Your AI runs the same add command you would.

Set up your AI

Pick your editor. Use the button or config below, or press Copy setup prompt and paste it into the editor's agent — it will make the change for you.

One command in your app's folder. --scope project writes .mcp.json so your team gets it too.

terminal
claude mcp add expokit --scope project -- npx -y @expokit/cli@latest mcp
Run this in your Expo app's folder. Starting the server by hand (npx @expokit/cli@latest mcp) prints nothing — it waits for your editor to connect.

Then just ask

Once connected, talk to it normally. A few prompts to start with:

Add a component

Add an ExpoKit bottom sheet to the profile screen and open it from the Edit button.

Build a screen

Build a sign-in screen with ExpoKit components: floating inputs, a password input and a primary button.

Find the right item

Which ExpoKit components could I use for onboarding? Show me their props before adding anything.

Swap in ExpoKit

Replace the hand-made toggle in settings.tsx with the ExpoKit switch and keep the same behaviour.

What your AI gets

  • list_items — Every component, block and template, with Free / Pro tags.
  • search_items — Find items by name or keyword ("sheet", "otp", "chart").
  • view_item — Props, import line, npm dependencies and usage examples. Free items can include their full source.
  • get_add_command — The exact add command for one or more items, and what it will change in your app.

Pro items

Your AI sees Pro items and their props, but not their source. To add them, run npx @expokit/cli@latest login once yourself (token from your /account page). Never paste the token into a chat.

AI without MCP

ChatGPT, Claude.ai and other chat apps can read the docs instead. Point them at /llms.txt (index) or /llms-full.txt (everything), or use Copy for AI on any component page.

prompt
Read https://expokit.karandev.in/llms.txt and use ExpoKit components for this Expo app. Add them with npx @expokit/cli@latest add <name>.
NextCLI

Did you like the content?