I've dabbled in building a tool of this type and something I've learned is that making single-task demo apps (like chess or snake game) is much faster and simpler than making production-grade apps that could make a real business. In fact, the process of making these demo apps production ready (polishing the UI, adding features, etc.) would usually result in the user giving up. In our implementation, it was much trickier for the system to understand the nuances of polishing than getting the broad strokes of the initial demo. I'd love to know if you've ran into this and what your thoughts are on getting around it.

Couple comments - have you played with Aider’s architecture mode? I think your workflows would benefit from it.

Right now it looks like the UI expert specs something and builds it, and then it asks for follow up.

I think the right dev flow working with a product dreamer involves UI mockups, then follow up questions, then implementation, with checkins — combining the idea of the architect with a “Product Manager” role is what most people need to get an app out.

Second thought is that I think you could see value with a bunch of custom prompted flows later in the app process: “do you want to take stripe payments?” “Do you want to add referral tracking support?” I’m not a native app dev, but I imagine there’s a bunch of CI, Firebase, other integrations that are best practice type things. Automating this as well would be really useful, and provide value and some lock in for your customers. To expand on this, you might make the basic option use your API access to these providers, and upcharge them. They could always implement their own if they want to move away, but the default path would get you some billing off every app that grows.

Another random idea would be ‘appifying’ a web site as a simple flow for the user. This feels like it might be even faster for the average person as a way to explain what they want, and would feed a lot of technical and visual direction straight into your workflow at the same time.

Anyway, good luck! I hope you guys get traction.

Glancing in more I see react-native mentioned, and expo+expo go mentioned, but like... it's been proven time and time again that Hacker News isn't your ideal feedback loop (ie: Dropbox, AirBNB).

Why make 4 paragraphs of text when you could just say:

- We obliterated X problem: (link)

- We demonstrate why that problem is significant here: (public link or private deck)

- We are growing at X->(timeframe or whatever impressive metric)

Reach out if interested. Thanks. [contact@info.com]

Recently came across this article where Gergely describes those 2 camps of "bootstrappers and iterators" - and that's basically what it is: v0/a0 for bootstrapping, cursor/copilot for actual coding https://newsletter.pragmaticengineer.com/p/how-ai-will-chang...

In particular, I'd hide/disable the sign-in-with-Google pop-up on phone-size screens (it takes a third of the screen for me), and make sure that CSS/styling makes everything fit on the code and preview screens on a phone, even in mWeb.

Can't get the QR Code to scan on my Pixel 8 in Expo Go. Camera just sits there staring at it. Could be unique to my phone not sure if working for others.

Got it to use an API I'm working with that requires JSONP to get around CORS. But this causes error to show about using browser APIS. Thought I could insert platform checks to determine web vs / android and whether to use JSON/JSONP but it still wants to remove any logic referring to web. Just FYI about the impedance mismatch of testing in browser assuming native environment.

I'm curious how commodotized the underlying LLM's are for generating this type of code? (And thereby, how big the moat is for companies like vercel/bolt surrounding v0 et al).

What models are you using to power the code editor? Did it require much customization?

Is it likely that enterprises would be able to eventually bring their own LLM's, or does the training make that prohibitive?

[1]https://www.ycombinator.com/launches/Mkl-onlook-cursor-for-d...

Is it possible to use Figma designs with this?

It's really hard to get something great through just a chat interface, and I've seen a lot of these AI tools don't allow me to actually give it more to work with like Figma designs.

Are you planning to offer your own backend, auth, notifications .. etc? Using Supabase will not keep users on your platform.

How do I share chat? I tried to build public bus tracker as a test but there are issues using external libraries.

Congrats on the launch Seth/Ayo!

It’s crazy to see that VCs companies would throw money for forking and copying an open-source project, and adjusting the prompt.

This kinda reminds me of Expo snacks (https://snack.expo.dev/), but with a chat.

Tons of work work to package everything in minimal decent functional product.

Great work, looking forward to use (and possibly subscribe) later.

I would imagine for those who want to are agencies / developers building apps who charge a fortune doesn't make sense anymore with tools like Replit.it, Bolt, Devin and now A0.

Great work Seth and Ayo for making it easier and potentially bringing to cost of building an app down close to free as I'm assuming this is now free as it is just a sign up.

Is there any pricing on this?

If I screenshot it, crop it and paste it into a blank doc, I can scan it.