Try Warp Build today and get an extra 1000 Oz credits ➞ https://oz.dev/betoyt

I connected my AI Tattoo app (Inkigo) to Warp's new cloud agent platform Oz, so every time a user submits a feature request, an AI agent spins up in the cloud, evaluates it, implements it, and opens a pull request — no local machine needed.

- Master React Native alongside me https://cwb.sh/rn
- Oz SDK: https://github.com/warpdotdev/oz-sdk-typescript
- Inkigo AI (AI Tattoo App): https://cwb.sh/inkigo

📬 Newsletter: https://cwb.sh/newsletter?r=yt
💬 Join the Discord community: https://cwb.sh/discord

For other ways to support my work, please consider:

- GitHub Sponsors: https://github.com/sponsors/betomoedano
- Become a Code with Beto channel member: https://cwb.sh/join

🐦 Follow Beto on X: https://x.com/betomoedano
💼 LinkedIn: https://www.linkedin.com/in/betomoedano
💻 GitHub: https://github.com/betomoedano
📸 Instagram: https://www.instagram.com/codewithbeto
🎵 TikTok: https://www.tiktok.com/@codewithbeto
🧵 Threads: https://www.threads.com/@codewithbeto
🦋 Bluesky: https://bsky.app/profile/codewithbeto.dev

⌚️ Timestamps ⌚️
00:00 Intro
00:31 The problem with local AI agents
00:54 What is Oz? Warp's cloud agent platform
01:16 Demo: AI Tattoo app (Inkigo) overview
01:53 Feature request form walkthrough
02:16 Plan: Triggering Oz from feature requests
02:32 Setting up Warp and the agent management panel
03:02 Testing a cloud agent
03:34 Creating an Oz environment for the project
04:48 Installing the Oz SDK
05:00 Using the Oz CLI
05:52 Connecting the SDK to the feature request API
07:02 Writing the trigger code
08:49 Configuring the model and base prompt
09:44 Live demo: Submitting two feature requests simultaneously
11:01 Monitoring parallel agents in real time
12:08 Reviewing the pull requests
12:42 Testing the pin prompt feature locally
13:33 Testing the camera grid feature locally
13:59 Follow-up prompt from the cloud agent
14:44 Shipping to production
15:01 Recap: What just happened + other Oz use cases

#ReactNative #AIAgents #WarpOz

Code with Beto

A developer demonstrates how to connect a live AI tattoo app to Warp's cloud coding agent platform (Oz) so that user-submitted feature requests automatically trigger cloud agents that evaluate, implement, and open pull requests — without any local machine involvement. The walkthrough covers creating an Oz environment linked to a GitHub repo, installing the Oz TypeScript SDK, and wiring it into an existing API endpoint. Two real feature requests are shown being processed in parallel, with working pull requests produced in minutes. The author also highlights other use cases like scheduled code cleanup, documentation updates, and Sentry error triage.

I Let My Users Build Their Own Features with AI Agents