pdepth

Inspiration

The inspiration behind pdepth cam from my own personal struggles with PDFs during my school days, too many PDFs and little time, so i decided to build a solution for my problem and that was how pdepth was birthed.

What it does

pdepth allows users to upload PDFs then gives them concise summaries and also recommend Youtube Videos for them to watch on site and solidify their knowledge, this is the first of its kind.

How we built it

pdepth was built using react and typescript for the frontend and fastapi for the backend, at its core, gemini and fireworks APIs were used for summary generation and some python libraries like PyMuPDF were used for PDF preprocessing and for the Youtube Video recommendations, an algorithm for contextual text extraction was used.

Challenges we ran into

During the building of pdepth I ran into tonnes of challenges but the following were the hardest of them, setting up the backend and building the best algorithm for the YouTube video recommendation.

Accomplishments that we're proud of

After using pdepth locally to suit my needs, i decided to let others use it and experience it, so i deployed it, deploying it was an acheivement on its own, but the feedback i recieved was really encouraging, as at now i have over 100 users and i just deployed it in september 2025.

What we learned

During the process of building pdepth i experienced a lot of challenges and i learnt a lot too, from building a web app from start to finish and handling user feedeback, just to mention a few.

What's next for pdepth

pdepth has a long way to go, i envision it having a million active users some day and also adding a lot of interactive features to keep users engaged.

Built With

• api
• fastapi
• fireworks
• gemini
• machine-learning
• python
• react
• typescript