Not Genius Team

Example output

Inspiration

Every day, millions of videos compete for attention, but most creators and small businesses still rely on guesswork when deciding what to post. While large brands can afford dedicated marketing teams and expensive analytics tools, smaller creators often cannot.

We wanted to build a tool that turns public trending data into practical growth decisions — helping users understand not only what is popular, but what is most relevant for their own product and audience.

What it does

TrendPilot is an AI-powered campaign planning tool that converts trending video signals into actionable short-form content strategies.

Users enter:

product or brand name
target audience
campaign goal
optional reference image

TrendPilot then generates:

multiple content angles
trend hooks
recommended posting times
thumbnail and editing guidance
audience engagement tactics
projected reach metrics

Instead of just showing trends, TrendPilot explains how to use them effectively.

How we built it

We built TrendPilot using:

Python + FastAPI for the backend
HTML/CSS + Jinja templates for the web interface
Pandas for data processing and scoring logic
OpenAI API for campaign generation and creative recommendations
GitHub + Vercel for collaboration and deployment

Our scoring pipeline combines trend momentum with business relevance, using weighted metrics such as engagement rate, freshness, publish-time growth, keyword match, and audience fit.

Challenges we ran into

One major challenge was balancing two different goals:

Trend popularity
Product relevance

Something highly viral is not always useful for a specific business. We had to design a system that blends both signals instead of over-prioritising one.

Another challenge was turning complex outputs into a clean and understandable interface within a short hackathon timeline.

Accomplishments that we're proud of

Built a working end-to-end web prototype
Successfully integrated AI-generated campaign outputs
Created a scoring system that combines virality and relevance
Delivered clear projections and creator-friendly recommendations
Produced a polished demo ready for real users to test

What we learned

We learned that raw data alone is not enough — users need interpretation, not just information.

We also learned how to rapidly prototype under time pressure, divide tasks effectively as a team, and turn an idea into a deployable product within a hackathon environment.