AiProPortrait

A year ago I thought the AI headshot wave was over.

Every week there was a new “LinkedIn photo in 5 minutes” startup on Product Hunt and X. As a solo dev, it felt like I’d already missed the boat.

But I kept noticing something:

People around me still didn’t have a professional photo they actually liked.

Friends hated their passport-style pictures. Hijab-wearing friends told me most AI tools either over-sexualized them, removed the hijab entirely, or produced something that just didn’t feel respectful. Many people didn’t have access to good photographers, or the prices were ridiculous compared to their salary.

So I built AiProPortrait.

Not because I wanted to ride the AI hype, but because getting a decent, professional headshot in 2025 is still:

• expensive

• awkward

• and for a lot of people, culturally off.

What I’m actually building

Right now AiProPortrait has two faces:

1. A Canva app (“Headshots”) used by designers, agencies, and regular people who live inside Canva.

2. A standalone web app for individuals and teams who want credits, galleries, and more control.

You upload casual photos → we train a custom model → you get a gallery of studio-quality headshots in outfits and backgrounds that look like something a real photographer could have delivered.

I’m bootstrapping and writing almost all the code myself:

• Next.js + React on the frontend

• Node/Express + SQL on the backend

• Stripe for payments / credits

• Auth.js for auth

• Diffusion models via third-party AI providers for training & generation

No templates, no white-label. Just a lot of evenings, coffee, and Tailwind classes.

We’ve crossed 500+ paying customers already and generate new headshots every day. It’s not “life-changing money” yet, but it feels like a real product that people rely on.

Why build this when there are bigger players?

Some reasons I’m still optimistic:

1. Niches inside niches

“AI headshots” sounds like a single product, but it’s really a bunch of micro-niches:

• corporate LinkedIn

• real estate agents

• modest / hijab-friendly headshots

• medical / residency (ERAS) photos

• actor / creator / dating profiles

• team photos for small companies

Most tools try to be a generic “AI avatar” thing. I’m focusing on cases where the photo has real stakes: job application, career branding, trust with clients.

2. Cultural sensitivity actually matters

If your product is only tested on people who look and dress like typical tech employees, you miss huge groups of users.

A lot of my users come from backgrounds that aren’t well represented in most AI datasets. I get constant feedback from customers who tell me “this is the first time an AI tool respected how I want to appear.”

That’s not just a “nice to have”. It’s a competitive advantage if I keep listening and designing for them on purpose.

3. Indie hackers can win with boring reliability

Most AI projects I see are demos with no support, no refund policy, no clear pricing.

My unfair advantages are small but real:

• clear credit-based pricing (no subscription needed)

• fast human support (I answer personally)

• refunds if results are bad

• continuous prompts/styling improvements based on real usage, not just model releases

It’s not flashy, but it builds trust.

The hard parts

It’s not all hockey-stick graphs:

• Infra costs and AI usage prices can eat your margin if you’re not disciplined.

• People expect magic from AI and can be disappointed when their selfies are low-quality or inconsistent.

• Support is emotional: people are using these photos to represent themselves professionally, so they care a lot.

The biggest challenge is balancing experimentation with stability. I want to try new models, new styles, and new flows, but breaking someone’s existing workflow for a “cool experiment” is not acceptable when they’re paying real money.

What I’m trying to learn from Indie Hackers

Here’s what I’m hoping to learn from the IH crowd:

• Stories from other founders selling AI-powered, credit-based tools. How do you price usage so you don’t get destroyed by heavy users?

• Whether anyone has successfully moved from a marketplace (like Canva) to a strong direct web customer base.

• How you’ve handled long-term dependence on AI providers (pricing changes, model deprecations, etc.).

If you’ve built something similar (AI product, design tool, or marketplace app), I’d love to hear your story.

Closing

I’m building AiProPortrait to give people a simple thing that somehow is still hard in 2025: a professional, respectful photo that they actually like and can afford.

If you’re curious, I’m happy to share numbers, experiments, or even the mistakes I’ve made with infra, CORS, and auth while growing this from a side project into a real business.

Happy to answer any questions or feedback from the IH community 🙌