Expo MCP Server: Connect AI assistants to your Expo projects
Key Points
- Remote MCP server for Expo projects
- Local SDK 54+ enables UI interaction & screenshots
- Preview available to EAS paid plan customers
Summary
The Expo MCP Server is a preview remote Model Context Protocol (MCP) server that connects AI coding assistants to your Expo projects. It gives assistants up-to-date Expo docs, intelligent dependency installation, configuration generation, and — with local capabilities enabled — visual verification and automated UI interactions against a running dev server.
Key Points
- What it is: A remote MCP server (mcp.expo.dev) that makes AI assistants Expo-aware and able to access project context and docs.
- Server capabilities: Always-available features include searching Expo docs, learning framework patterns, generating configs, and recommending/installing compatible packages (npx expo install).
- Local capabilities (SDK 54+): When your Expo dev server runs with MCP enabled the assistant can capture screenshots, simulate taps, open DevTools, and verify UI behavior.
- Use cases: Automate dependency management, implement features from docs, run visual verification and fix UI issues autonomously, and augment testing (lint/type/unit) with visual checks.
- Requirements: Any EAS paid plan (hosted MCP available to customers), Expo SDK 54+ for local features, and AI tools that support remote MCP (e.g., Claude Code, Cursor, VS Code extensions).
- Preview notes: Feature is in preview; Expo requests feedback via Discord (#mcp) and Twitter/X to prioritize future EAS integrations (builds, updates, app store automation).
Getting started (practical)
- Ensure you have an EAS paid plan and Expo SDK >=54 for local features.
- Configure MCP in your project (takes ~1 minute) and connect your preferred AI tool that supports remote MCP.
- Start your Expo dev server with MCP enabled to unlock local automation and visual verification.
Recommended next steps for engineers
- Try common workflows: package installs for your SDK, adding camera/notifications, and visual tests (screenshots + interaction checks).
- Provide feedback on desired EAS integrations and automated verifications to shape SDK 55+ roadmap.