I'm a student, and I have a pretty simple problem: I can't write fast enough to keep up with my classes.
I'd come home from lectures with half-finished notes, gaps I couldn't fill in, and the sinking feeling that I'd missed something important. I tried voice recorders, I tried apps, I tried just… accepting it. None of it really worked the way I wanted.
What I actually wanted was simple: something that records automatically, transcribes what's being said in real time, and then gives me a clean summary I can actually study from. I didn't want to babysit it. I just wanted to show up, hit record, and have useful notes waiting for me afterwards.
So I built it.
It's called Notetaker. It runs on macOS, it transcribes your audio live using Apple's on-device speech recognition (so nothing leaves your computer), and when you're done it uses an AI model of your choice to generate a summary. You can use a local model like Ollama if you don't want anything going to the cloud at all, or plug in OpenAI or Anthropic if you prefer. I set mine up to just use Apple's built-in foundation model by default — no setup needed, totally private.
You can also schedule recordings in advance, which is the part I use most. I just set it to start when my class begins and forget about it.
It's not a polished commercial product. It's something I made for myself that I think other students (or really anyone who sits through a lot of meetings or lectures) might find useful. The code is on GitHub and it's free.
If you find it helpful, buying me a coffee helps me keep working on it. But honestly, just knowing someone else is using it is already pretty cool.
Thanks for reading — Steven