Context: it generates images with early 2000s digital horror aesthetics - think Blair Witch Project stills, early creepypasta, that specific low-res dread from when internet horror was new. No login, completely free, uses Gemini for prompt enhancement and Replicate for generation.
What I'm seeing: - People seem to prefer constrained tools for specific use cases - "Generate anything" tools are overwhelming for most users - Niche aesthetics can't be easily prompted in general-purpose models - Artists want consistency, not infinite options
But the economics are weird. API costs mean I can't really monetize without adding the exact friction (logins, credits) that makes the experience worse. The value is in it being immediate and free, but that's not a business model.
Questions: 1. Do specialized aesthetic generators have actual staying power, or is this just novelty? 2. How would you monetize something like this without ruining the UX? 3. Is there technical merit in building domain-specific image models vs. prompt engineering general ones?
The technical side interests me more than the business side - getting period-specific aesthetics right required digging into camera sensor specs from 2001, JPEG compression algorithms, CRT color profiles. That knowledge feels weirdly valuable but I don't know what to do with it.
Current cost is maybe $50/month in API calls. If it 10x I'd have to change something, but I also don't want to kill it by adding payment walls.
Anyone else built niche creative tools? What happened?
Link if you want to try it: https://dreamyy2k.app