Thanks for the thoughtful and critical feedback, this is exactly why I posted here.
You raised fair points about the mixed messaging, and I’ve just pushed updates to address them:
• Privacy Policy & data collection: You're right that the tagline “we never collect data” was too absolute. I do use standard analytics (GA) for anonymous usage metrics and error tracking. The Privacy Policy now clearly separates File Data — which is processed 100% locally and never leaves the browser — from Usage Metadata, which is anonymized and collected only for understanding feature performance.
• Network activity: The POST requests you saw come solely from those analytics libraries. No file contents, pasted text, or conversion results ever hit the network. I’ll also review whether I can reduce or defer analytics calls to make this more transparent.
• Visibility of terms: Agreed. I’ve added a prominent Privacy/Terms link in the header and a first-visit consent banner so users aren’t relying on a tiny footer link or assumptions.
• Offline behavior: The conversion logic runs entirely in Web Workers and doesn’t require a server, but my PWA config wasn’t robust enough to guarantee a clean offline startup. I’m working on tightening that up so users can verify the “local-only” behavior themselves.
None of this was intended to be sketchy — I simply oversimplified the marketing copy and didn’t surface the right information. I really appreciate you calling it out and giving me the chance to improve it.
I just tested the Gemini 3 preview as well, and its capabilities are honestly surprising.
As an experiment I asked it to recreate a small slice of Zelda , nothing fancy, just a mock interface and a very rough combat scene. It managed to put together a pretty convincing UI using only SVG, and even wired up some simple interactions.
It’s obviously nowhere near a real game, but the fact that it can structure and render something that coherent from a single prompt is kind of wild. Curious to see how far this generation can actually go once the tooling matures.
Mini-apps feel like a way for major platforms to control more user entry points. They sit between web and native apps and are much faster to build. From my perspective, they’re just another iteration of how platforms experiment with lighter app formats.
AI has lowered the barrier for building apps so much that even small teams can ship useful mini-apps quickly. A policy shift like this will probably pull more developers into the ecosystem, which usually leads to more experimentation and a healthier, more diverse set of apps overall.
Developing on Windows can be troublesome at times, but I still prefer how it feels to use.
For development alone, macOS is probably the best choice since it provides a native Unix environment.
I just happen to enjoy the Windows interface more, even if it means dealing with extra setup.
China’s AI scene is moving fast, and this kind of visa could make it move even faster. But it’s been a hot topic back home , a lot of people are supportive, while others are quite skeptical.