AI Tool Integrations (MCP)
Aurora works as an MCP server, so you can build and deploy projects directly from your favorite AI tools.MCP Integrations
Use Aurora from Claude, Claude Code, Cursor, Windsurf, and any MCP-compatible tool. Just paste a URL and authorize — no config files or API keys needed.
Native Integrations
GitHub
Import existing repositories (including Lovable projects), automatic repository creation, commit syncing, and full version history under your own GitHub account.
Supabase
Authentication, Postgres database, storage, and Edge Functions, connected directly to your project.
Stripe
Add one-time payments, subscriptions, checkout flows, and billing portals with Stripe and Supabase Edge Functions.
Using Other APIs
Aurora can work with any REST API. Your projects run in full Node.js containers, so you can install any npm package and call any endpoint.What works well
- REST APIs with clear documentation
- APIs with OpenAPI/Swagger specs
- Services using API key or Bearer token authentication
- Any npm package (Stripe SDK, OpenAI SDK, etc.)
Tips
- Install the SDK: Ask Aurora to add the package (e.g. “add Stripe payments”) and it will install the npm package and write the integration code.
-
Store secrets safely: Add API keys in Settings > Secrets. They are encrypted and injected as environment variables. Read private keys in server code with
process.env.YOUR_KEY. OnlyVITE_-prefixed variables reach browser code, throughimport.meta.env.VITE_YOUR_KEY. - Provide context: Share API documentation or examples in the chat. The more detail you give, the better the integration code will be.
- Start simple: Begin with read-only endpoints before attempting complex workflows.
Need help with a specific integration? Just ask in the chat. Aurora can install packages, write API calls, and handle error cases for most popular services.