Sabitlenmiş Tweet

To celebrate @cluely's $15M round by @a16z, we're launching the same thing... for free! Oh, actually, we're completely open-sourcing it too.
Good for them on the marketing front. In fact, despite the criticisms, they're pushing boundaries for one thing. I had always thought of overlays but never implemented them, so it definitely stirred me to act.
So now's the time for us to push further too :)
The core logic is the same as Cluely's:
- Take a screenshot via the Apple SDK
- Transcribe audio in parallel (not added yet)
- Feed it to GPT
- Stream it back
We went even a step beyond and figured out how to do instant OCR in 200ms to generate even better responses. Screen parsing + image.
I haven't added audio yet, as I don't attend meetings and have been busy building something else.
However, I'm sure that if someone can Venmo me $300, we can create an open-source bounty to add it in.
I'll add the links in the comment below, check them out, and I'm pumped to be alive at this time in history.
English

















