UploadKit

• 

  Start

• 

  Put transcript and press Generate

• 

  It will suggest title name and description

Inspiration

Uploading a video is more than hitting publish. A good title, description, tags, and chapter list all take real time to write, and most people either rush them or skip them entirely. I was already building AI tools for another hackathon and wanted to try automating the exact part of the creator workflow that eats the most time without adding much value.

What it does

You paste in a transcript, a topic, or just rough notes about your video, and it generates five title options, a ready to paste description, a list of tags, and chapter timestamps when your notes have timing info to work from. Every section has a one click copy button.

How I built it

Next.js and Tailwind CSS on the frontend. One backend route sends your input to an AI model and asks for everything back as a single structured response, which the page then renders as titles, a description block, a tag list, and chapters.

Challenges I ran into

Making sure chapters only show up when the input actually has timing information, instead of letting the model invent fake timestamps, took some careful prompt writing. Getting reliable, cleanly parseable output back from the model every single time took a few rounds of tightening the prompt too.

What I learned

A lot of this carried over from a different AI project I built earlier, mainly how to prompt a model for structured output you can actually parse and render, and how to scope a tool around one real, specific workflow step instead of trying to do everything.

What is next

Batch processing so you can generate kits for multiple videos at once, and maybe a browser extension version that fills the fields in directly on YouTube's own upload page.

Built With

• deepseek-api
• next.js
• nod
• react
• tailwind-css
• typescript