Whisp is a Chrome extension that brings voice dictation anywhere in the browser.
Q: Why not just use Apple's Dictation?
OpenAI's Whisper model is a lot better.
It can accurately recognise my voice when I'm (1) whispering at the library or (2) in a crowded environment like a loud coffee shop.
Q: Why not use SuperWhisper or Whispering?
SuperWhisper is paid, and it doesn't justify the cost for my use case. I'm primarily interested in the transcription part and the API cost is almost nothing.
I tried Whispering but I couldn't get it to run in the browser.
As always, I think there's room for improvements to the UX in ways that are specific to my own use case.
I decided to develop this extension with the following requirements in mind:
-
It should feel native to the browser that I'm using (Arc) The UI should be unobtrusive but feel responsive and natural. When activating the extension, it should start recording immediately.
-
It must feel seamless for the main use case - building prompts When talking to ChatGPT or Claude, I have a tendency to write my prompts in this format:
< Some context explaining what I'm doing > < Pasted code / homework assignment / essay question etc > < Explanation of my approach to the problem > < Pasted answer to the question > < Instructing ChatGPT / Claude to do X task >
As I'm pasting some text into the chat from some other source, I can verbally describe what I want the AI to do with it. The final output should appropriately interleave the voice and pasted text segments.
Voice -> sending new prompt