SwiftAI Boilerplate Pro
Launch an AI-powered iOS app in a week or less
10 followers
Launch an AI-powered iOS app in a week or less
10 followers
SwiftAI Boilerplate Pro helps indie hackers, solopreneurs, and iOS developers build AI-powered apps faster. It comes with production-ready auth, AI chat UI, subscriptions, and SwiftData. It is battle tested with Capishi, an AI interview app I shipped in under a week using this exact boilerplate. Built for builders who want to move fast without sacrificing clean architecture.
Hello Product Hunt!
I'm Berkin, an iOS engineer and indie hacker building AI-powered products.
I've had a bunch of ideas for AI apps, but it was hard to decide which one to start with. As I sketched them out, I realized most of them shared the same core needs: login, subscriptions, AI chat UI, data storage, and authentication, and I kept rewriting the same code.
So instead of building one app, I built a solid boilerplate to support them all.
That became SwiftAI Boilerplate Pro, and I used it to launch Capishi, an AI interview coach app, in less than a week.
The boilerplate includes:
• Supabase auth (email, Apple, Google)
• RevenueCat subscriptions
• AI chat with OpenRouter (supports both streaming and non-streaming requests)
• SwiftData and Keychain for storage
• OneSignal push notifications
• Clean MVVM architecture with protocol-based DI
It's already made a few sales (almost $400), and I've been in contact with every customer. One of them asked for push notifications, so I added OneSignal. I'm building this in public and will keep improving it based on real feedback.
Most customers so far are indie devs and solo founders, though one iOS engineer is already using it at work. It's built for speed but grounded in clean, scalable architecture, perfect for early-stage apps.
I'd love your input:
* What kind of AI app would you build with this?
* Would Firebase support (alongside Supabase) be useful?
* What other features or integrations would you want?
Thanks for checking it out. I'll be here all day to chat, answer questions, and hear what you're working on too.
Love that you're building in public based on user feedback. How easy is it to swap out OpenRouter for another LLM provider, or adjust the chat UI for non-AI features?
@berkin_sili
@olajiggy321 Thanks for the thoughtful question and glad that you liked the idea! I designed the architecture to make both of these things as easy as possible.
Swapping out OpenRouter is very straightforward. The AI layer is protocol-based, so any provider can plug into the same interface. If you want to swap out OpenRouter, you just add a new LLMClient implementation and update a single line in the Composition Root. The Supabase Edge Function can also be switched by changing the API endpoint and headers; it's about 150 lines of TypeScript, so you could point it at OpenAI, Anthropic, or a self-hosted model in under 10 minutes.
The chat UI is fully modular as well. Each part, bubbles, layout, streaming bubble, input bar, starter prompts, is its own component. You can restyle them, change the layout, or even replace the bubbles with cards, images, or video previews. A few buyers are already adapting it to different use cases like journaling and scriptwriting tools. And if you want to experiment with other or private LLMs, I'm more than happy to help out.
If you want to explore any of these parts in more detail, I'm happy to walk you through how it's structured.
@berkin_sili
Thanks for the detailed answer—modular architecture and easy LLM swaps make this really flexible.
I have a small idea related to improving the flexibility of those chat components that you could try on your own.
If you're open to a suggestion, what's the best way to share it?
@olajiggy321 Thanks a lot, I appreciate that. I'm always open to suggestions. Feel free to share it here in the thread, or DM me on X (@berkinsili), happy to take a look.
@berkinsili @berkin_sili
kindly folow back on X so i Can dm thanks