GETTING STARTED
Init, configure, and deploy your first runtime.
Cheela ships as a single CLI plus a thin SDK. Five steps from init to your first deployed runtime.
1. Initialize your project
$ npx cheela init
✓ created cheela.config.ts
2. Add your API key
// cheela.config.ts
export default {
apiKey: "sk-live-••••••••",
};
3. Write a capability
export const summarize = defineCapability({
input: "text",
output: "summary",
});
4. Create a runtime
$ npx cheela runtime register \
--provider openai \
--model gpt-4.1
✓ runtime rt_8f2a registered
5. Deploy
$ npx cheela deploy
✓ runtime rt_8f2a deployed · v1
That's it — routing, retries, and eval scoring run automatically, and every execution shows up in your dashboard under Executions.
Want to try this without installing anything?
Run it live in the playground.