FileGPT

Drop in a scanned PDF, a YouTube link or an audio recording, then ask a plain question and get an answer pulled from that file instead of scrolling through it yourself.

That is the basic job of FileGPT, an AI chatbot and document extraction tool that lets you chat with your files in natural language.

FileGPT accepts several kinds of material: text files, YouTube videos, audio, webpages and PDFs.

Scanned documents are handled with OCR, so a PDF that is really just an image of a page can still be searched by question.

Everything you upload becomes a personal knowledge base you can query, which suits students reading long papers, people summarizing lectures or videos, and anyone who has to dig through reports for one figure.

The app runs on GPT models, so answers depend on what the model can find in your files.

See it as a faster way to locate and summarize information, and check important details against the source before you rely on them.