3
4 Comments

How I Built PhotoProAI in 3 Days and Got 784 Signups and 10 Paid Users in 2 Weeks

Hi, indie hackers. I'll share my story about this topic.

Introduction:

Two weeks ago, I launched my new product PhotoProAI after building it in just 3 days. Since then, it has gotten 784 signups, 10 paying users, and users have generated 344 pictures with it. In this post, I'll share the story behind the idea, how I built it so quickly, and how I got initial traction.

The Idea Origin:

I found that taking high quality headshot photos was incredibly time consuming and tiring. Every time I wanted a nice headshot, I'd have to meticulously set up the lighting, carefully pick my clothes and makeup, and take dozens of shots to get one good one.

Even after taking the photos, the post processing work was grueling. I'd spend hours painstakingly Photoshopping each photo to adjust the lighting, remove blemishes, tweak the facial features, apply digital makeup, and more. All that effort for just one nice headshot!

I realized there must be a better way. With the rapid advances in AI generation, I wondered if AI could help simplify and automate this process. That sparked the idea for PhotoProAI - a web app where users could easily upload any headshot photo and instantly get back enhanced and beautified versions.

What makes PhotoProAI unique is it only requires a single photo, unlike competitors that need 20+ photos. This lowers the barrier for getting great results.

Building in 3 Days:

I built PhotoProAI rapidly over 3 days using:

  • Next.js for the frontend React framework - this let me use React best practices while optimizing for SEO and performance right away.
  • Tailwind CSS for quick styling - I could style pages quickly without writing custom CSS.
  • Stable Diffusion hosted on a Python FastAPI server for the AI model. FastAPI made it simple to set up REST API endpoints to generate images.
  • Celery and RabbitMQ to queue up generation requests and prevent blocking the main thread. This improved user experience.

Getting Initial Traction:

To get those initial 784 signups after launch, I focused on three strategies:

  • Engaging with relevant communities on Reddit, Indie Hackers, and Twitter. I introduced the product and shared my journey to spark interest.
  • Reaching out to friends directly to try out the product and give feedback. Their thoughts helped shape the initial product direction.
  • Responding quickly to user feedback to fix pain points and improve the onboarding experience. This increased signups and conversions.
posted to Icon for group Ideas and Validation
Ideas and Validation
on July 27, 2023
  1. 1

    I am a web develop with next js and node js please tell what I have to learn new for that
    Does need generative or other technical skills or everything is done from resources

    1. 1

      You need to learn some stable diffusion knowledge.

  2. 1

    how you train that diffusion model

    1. 1

      I used koyha_ss script to train the lora model.

Trending on Indie Hackers
I built an AI that turns an idea into a live business in under 10 minutes. Here’s what 1,000 launches taught me User Avatar 76 comments I built a web-based vector editor from scratch and integrated an AI Agent. Need just ONE beta tester! User Avatar 70 comments Building Noodle, a keyboard-first REST client for the terminal User Avatar 31 comments "Looks Good to Me" Is Quietly Killing Your Feedback Loop User Avatar 28 comments Launched 580 landing pages in 1 week. Solo. No team. User Avatar 18 comments I built a competitor monitor for indie founders User Avatar 14 comments