Inspiration

I was inspired to make this project because of the insane amount of time it often takes to create flashcards when the content you are studying isn't in text form.

What it does

It allows you, with the click of a button, to listen to any audio on the chrome browser, and then generate flashcards based on a provided prompt which can then be exported to csv or Anki.

How we built it

I built this in the form of a Chrome extension and Chrome storage, audio transcription via the ElevenLabs API, and processing through Gemini.

Challenges we ran into

A big challenge that I ran into is that the Gemini API was initially very slow to process even small inputs. I was able to fix this via disabling all but the most necessary features.

Accomplishments that we're proud of

I am proud of the level of customization possible here. You can use this extension to study virtually anything: CS, Biology, languages, etc.

What we learned

I have learned that LLMs, while powerful, still have their limitations. You can't throw an infinite amount of work at them and expect everything to just magically function perfectly.

What's next for AnkiCapture

I think that the next logical step, if I were to publish it, would be to create a small backend to have a bit more security in user-provided API keys. Apart from that, adding some non-audio features might be helptful.