AdBrain

Homepage
Logo
Preview Page

Inspiration

In today’s creator economy, businesses, especially startups need high-quality video ads every day, but scripting, recording, and editing can take hours of time and money that these smaller companies may not have. With the rise of AI video generation tools like Sora and Veo, we realized we could automate this entire process and increase the quantity of videos outputted thus increasing total user engagement. Our goal: empower entrepreneurs, e-commerce sellers, and marketers to turn simple text ideas into viral-ready ad videos in minutes.

What it does

AdBrain is an AI-powered ad generator that transforms product ideas or short prompts into ready-to-post ads.

Here’s how it works:

Users enter a short creative prompt or upload a product image.
Our backend generates a short-form video using advanced text-to-video models.
AdBrain automatically adds motion, captions, and audio designed for engagement.
Finally, with one click, users can post directly to Instagram Reels using our integrated Meta Graph API.
Users can view their generated ads on their personal dashboard.

Within minutes, a fully polished, platform-ready ad is live.

How we built it

Frontend: Next.js + Tailwind for a clean, responsive dashboard (prompt input, video previews, user dashboard, upload status).
Backend: FastAPI (Python) handles prompt requests, video generation, and database operations.
Database: Supabase (PostgreSQL + Auth) for user management, video metadata, and storage linking.
Storage: AWS S3 for raw and generated video files.
Video: Integrated text-to-video pipeline using Google Veo model for video generation with FFMpeg for concatenation,
Social Media Integration: Meta Developer API for posting generated videos directly to Instagram Reels.
Async Infrastructure: Background tasks and presigned URLs for smooth uploads and minimal latency.

Challenges we ran into

Handling video upload reliability (FFmpeg parsing and presigned URL edge cases).
Managing API rate limits and async background processing for the Veo model since it is expensive + slow.

Accomplishments that we're proud of

Fully working end-to-end pipeline: prompt → video → Instagram Reel
Successfully integrated AWS S3, Supabase, and FastAPI into one cohesive system.
Created something that’s both technically deep and commercially impactful a real product, not just a demo.

What we learned

How to design robust asynchronous backend services for large media workflows.
How to balance AI creativity with user control (prompt templating, content safety, and video aesthetics).
Collaboration under tight deadlines: delegating tasks across model pipelines, UI, and infra.
The importance of early architecture planning and the power of clean abstractions.

What's next for AdBrain

Expand to multi-platform posting (TikTok, YouTube Shorts).
Integrate AI-generated voiceovers, captions, and thumbnails.
Add trend analysis + product keyword generation for e-commerce users.
Launch a public beta where creators can monetize and customize generated ads.
Deploy a fully managed cloud infrastructure on AWS Lambda + Supabase Edge Functions for scale.

AdBrain’s mission is to give every creator and brand a creative partner that never sleeps.