AI Headshots
AI Headshots
AI Headshots
Renas Hassan
Oct 15th, 2025
Overview
Getting a professional headshot usually means booking a photographer, finding a studio, picking outfits, and spending a few hundred dollars for a handful of photos. It's a process most people put off because of the cost and effort involved.
I built AI Headshots to solve that. It's a platform where users upload a few selfies, and the AI trains a custom model on their face to generate realistic, studio-quality headshots in minutes. No photographer, no studio, no hassle.
The result is a set of professional headshots with different outfits, backgrounds, and poses that look indistinguishable from photos taken in a real studio.
The problem & motivation
A professional headshot is one of the first things people notice on your LinkedIn, resume, or company website. It signals credibility and makes a strong first impression.
But the traditional route is painful. A studio photoshoot costs $250+ and takes days to get results back. Most people either skip it entirely or use a low-quality selfie instead.
I saw an opportunity to make this accessible to everyone. With advances in AI image generation, it's now possible to train a custom model on someone's photos and produce results that rival professional photography, at a fraction of the cost and time.
Key features & functionality
- Selfie-to-headshot transformation: Upload a few selfies and receive dozens of professional headshots with various outfits and backgrounds
- Custom AI model training: Each user gets a personalized AI model trained on their specific features for realistic results
- Fast turnaround: Results delivered in under an hour, with premium plans ready in under 30 minutes
- Multiple variations: Up to 200 headshots with different poses, outfits, and backgrounds per session
- Multi-step upload flow: Guided process with photo guidelines to ensure the best possible results
- User dashboard: Browse, manage, and download generated headshots
- Stripe billing: Three pricing tiers with secure payment processing
- GDPR-compliant: All uploaded photos deleted after 7 days, generated headshots deleted after 30 days
Tech stack
- Core: Next.js, TypeScript
- Frontend: React, Tailwind CSS, shadcn ui, React Query
- Backend: tRPC, Drizzle ORM
- Database: PostgreSQL
- AI: FAL AI for model training and headshot generation
- Payments: Stripe
- Storage: AWS S3
- Auth: NextAuth.js
- Email: Resend
- Monitoring: Sentry
- Code Quality: ESLint, Prettier, Husky
Challenges & learnings
AI model quality and consistency
The biggest challenge was getting consistently good results from the AI. Not all selfies produce great headshots, so I had to build a guided upload flow with clear photo guidelines to help users submit the right kind of images. Things like lighting, angle, accessories, and photo clarity all affect the output quality significantly.
I also had to experiment with different training parameters and prompts to find the sweet spot between realism and variety. Too little training and the headshots don't look like the person. Too much and you lose diversity in the outputs.
Async processing and user experience
Since training an AI model takes time, the entire generation pipeline is asynchronous. Users upload their photos, the model gets trained in the background, and then headshots are generated and delivered. Managing this flow while keeping the user informed about progress was a key UX challenge.
I used webhooks from the AI provider to track generation status and update the user's dashboard in real time as headshots become available.
Privacy and trust
When you're asking people to upload photos of their face, trust is everything. I put a lot of effort into making the privacy policy and data handling transparent. All uploaded photos are automatically deleted after 7 days, and generated headshots are removed after 30 days. Users can also manually delete their data at any time from their account settings.
Being based in Sweden and operating under GDPR gave users additional confidence in how their data is handled.
Outcome
The platform is live at aiheadshots.se and has received positive feedback from users who appreciate the quality of the results and how quick the process is compared to a traditional photoshoot. The generated headshots are realistic enough that people use them on LinkedIn, resumes, company websites, and email signatures.
Building a product around AI image generation taught me a lot about managing async workflows, working with external AI APIs, and the importance of user trust when handling sensitive data like personal photos.