Inspiration

I really like the application of AI image generation into advertisements because in india, Sometimes Instagram algorithm can really show you some terrible ad designs especially by SMBs who don't have high budgets for ads, That's why I wanted to create something like Adcraft AI, so that even small business owners with no marketing budget can transform their product images to stunning ads within seconds.

I wanted it to be felt intuitive and empowering, allowing users to transform a simple product image into a compelling ad or generate entirely new visuals from a textual idea. The idea was to bridge the gap between a raw concept and a polished, professional advertisement, all within a single, user-friendly interface. You just have to upload your product image and Adcraft AI does all the work like creativity, art direction, ad concepts etc.

What it does

Adcraft AI is a web app that helps you create awesome product ads within seconds.

You upload a product photo, and AI instantly turns it into a professional ad image. Then, you can edit it right on the screen – add text, shapes, or emojis, or even remove the background of your product with one click. You can also generate completely new images from text using AI. Finally, you export your finished ad to share.

It's your easy, all-in-one AI ad studio.

How I built it

I was building a complete creative ad agency as my main project for this hackathon then after running into some problems I decided to switch because of lack of time and build the main features with single prompt in one-shot challenge.

From experimenting with my main project I already had some rough Idea of how to implement features I want within bolt like:

  • I knew how to integrate fal ai with flux and openai api.
  • major things I had figure out was how to implement canvas edit functionality for
  • Ad Generation flow: image go to openai model -> we receive an Ad generation prompt -> this prompt + original image will be sent to flux/kontext on fal ai -> we receive ai generated ad.
  • Image Generation: we can also generate new images with flux/schnell on fal ai.

Challenges we ran into

  • Main challenge was to integrate the canvas functionality and make it work with AI image generation.
  • I prioritized UI very much, so deciding on theme, color scheme, style was also major part of this project. I used fabric.js for the canvas implementation.
  • had to figure out how to edit ai images as i don't have access to open ai imagen 4, so I integrated two llms together open ai for creative ad prompt and flux/kontext for editing the image

Accomplishments that I am proud of

Really worked so hard to complete this mostly because I felt defeated as I couldn't completed my main project. As I had no time left it felt like this is my last chance, but it was all worth it by seeing the one shot result from bolt. It really surreal how advanced tech have become, thanks bolt team!

Built With

Share this project:

Updates