I built a tool to generate user personas from a product idea. Soft-launched it this week.
The idea is simple — describe what you're building, get around 3 detailed user personas in about a minute. Demographics, pain points, motivations, behaviors. Export to PDF or JSON. That's it.
I built it because every time I started a new product, I'd do the same embarrassing thing. Open a blank doc, write "User Personas" at the top, and describe myself with different names. Turns out almost everyone does this. And it's a big reason products miss.
Proper persona research takes 4–8 weeks and costs $5–50K if you do it right. So 95% of teams skip it entirely — or fake it and hope for the best.
I'm not a full-time founder. Day job in tech, building this on nights and weekends. It's live now. Free to try. First persona set is on me.
If anyone wants to roast it: https://personaforge.pehloo.xyz/
Brutal feedback welcome. Especially:
- Did the personas actually feel useful or just generic AI slop?
- Is the input flow confusing?
- Would you pay for this, and if so what would make you?
Cool product. The blank doc with user personas at the top and describing yourself with different names line is too real. I have been doing persona work for something I am building and it is genuinely hard to do well without falling into that trap. Definitely going to try it out.
I know a few early-stage startup founders who skip formal persona research and are struggling with user understanding. I'm sure they'd be willing to answer your questions for free; happy to forward them if you'd like.
The utility is clear.
The trust gap is the real product.
This category dies fast if the output feels like polished AI slop pretending to be research.
Users can tell in 30 seconds whether this is insight or just generated formatting.
So the value is not “3 personas in a minute.”
It’s whether the output gets uncomfortably specific enough to surface the real buying objections, false assumptions, and demand risks that actually kill products.
That’s the layer people would pay for.
Right now Personaforge explains the function, but it still reads more like a prompt utility than a serious research product.
If you keep building it, the product likely needs to feel less like persona generation and more like conviction-grade market intelligence.
Beryxa.com would carry that much better if this becomes more than a side tool.
Appreciate the feedback! Yeah its more slop-y at the moment, but adding a background research agent to help improve the personas which should start surfacing the real things to poke into.
That’s the right direction.
The moment it starts surfacing things users did not already know, it stops feeling like persona generation and starts feeling like actual market signal.
That’s the line that matters here.
Until then it gets judged like a prompt tool.
After that it gets judged like research.
That’s also where the current name starts getting expensive fast.
Interesting idea — persona creation is one of those things everyone knows they should do but rarely gets done properly.
I’ve noticed most early-stage founders don’t fail because of ideas, but because their “assumed user” is basically themselves in disguise.
Curious — how are you validating that the personas generated actually match real user behavior, not just AI patterns?
That's a great question. And, well unlike many other tasks there is no benchmark for this. The only "real" way is if someone launches a product that actually works due to these personas. I'm building this out using this tool itself, so if it works then that's the best validation.
Apart from that I do have a few basic evals, but maybe a little more time on it won't hurt. Thanks!
This is actually super relatable 😄
The “describe yourself with different names” part is way too real — I think most people do that without realizing it.
The speed + low friction angle is strong here. Getting usable personas in a minute is a big unlock compared to weeks of research.
One thing I’d watch though:
→ how do you keep them from feeling generic over time?
Maybe adding:
→ industry-specific context
→ or pulling signals from real data (URL, product, audience)
could make them feel more “grounded” vs AI-generated.
Also curious:
Have you seen people actually use these personas in decisions (copy, features, etc.), or mostly just generate and leave?
Overall, solid direction 👍
Yeah that's the next step for me. I'll add a do my homework mode which does the thing of running a background agent to gather signals from real data.
Right now yeah there's not much retention, as its just people trying it out for fun. So I need to add some "real" features, and hence looking for feedback. Appreciate your thoughts!
Hey Ribhu, love the niche. Persona research is definitely a massive 'blank page' problem for founders.
I ran your site through my tool, RoastMyLanding, to answer your question: 'Is the input flow confusing?'
Your Trust Leak Report: 🔗 https://roastmylanding.vercel.app/roast/raQC456_
Would I pay? I’d pay if the export felt 'investor-ready.'
I’m doing a launch special—the full 10-point roadmap is currently 50% off ($9.50) with code TRUSTLEAK50 if you want to see the exact UI/Copy fixes needed to convert those 'soft-launch' visitors into users! 🚀