Inspiration

Content teams juggle fragmented tools and repetitive edits

What it does

Faster asset creation, consistent quality, lower switch costs

How we built it

Frontend: The frontend work included the user interface, where the user can use the tool to upload their file and enter suggestions. Languages used were HTML, CSS, and JavaScript. A simple interface was created using HTML & CSS, and JavaScript was used to connect the frontend to the backend as well as give functionality to the interface.

Backend : Developed a modular backend for AdMuse, integrating multiple AI-driven agents — Director, Variants, Scoring, Compliance, and Packager — to automate creative generation. Implemented Fetch.AI integration for intelligent scoring and optimization, added .env configuration for secure API key management, and used Pillow for creative rendering and export packaging. Tested and deployed locally on Python 3.12 (Windows) using virtual environments and version-controlled via GitHub.

The presentation was made using Figma and Powerpoint, and the code can be accessed on Github.

TABLE 144 A Harshitha Mallappa, Ibrahim Tahir, Prabhleen Gill

Share this project:

Updates