# Build it with AI

Copy one prompt into Claude Code, Cursor or ChatGPT and your coding agent wires Apexly into your codebase end to end, from the customer's hostname to webhooks and signature checks.

Paste this into a coding agent opened in your repository and it builds the whole integration with you: a server-only API client, the customer's "add your domain" screen, the webhook handler and the origin signature check. It reads these docs through [llms.txt](/llms.txt) as it goes, so it uses current endpoints, not a guess.



The prompt in your dashboard already has your app id, origin and API base filled in, and includes a freshly created test key. Find it on the first-domain screen and in **Settings**, under Signing secret.

## Which tool

- **Claude Code or Cursor:** open the agent in your repository and paste. It reads your code, tells you its plan, then makes the changes.
- **ChatGPT or another chat:** paste, then attach the files it asks for (your router or middleware, the tenant model, the settings page). Copy its changes back by hand.
- **Want the agent to manage domains too?** Add the [Apexly MCP server](/docs/guides/mcp) so it can create and check test domains while it builds.

## Keep your keys safe

- Use a **test key** while building: test domains never contact a certificate authority and walk every state on a ten-second timer. See [Test mode](/docs/guides/test-mode).
- The dashboard only ever puts a test key, never a live one, into the prompt. Live keys stay in `APEXLY_API_KEY` in your environment or secrets manager.
- The signing secret is never in the prompt. The agent reads it from `APEXLY_SIGNING_SECRET`.