Hacker News new | ask | show | jobs
I Made a ChatGPT Microphone Chrome Extension (Open Source) (chrome.google.com)
5 points by david-venegas 1139 days ago
2 comments

TL;DR: Created a Chrome extension called ChatGPT Microphone for voice-to-text with ChatGPT, featuring hotkeys, shortcut snippets, and a seamless design. It's free, ad-free, and open-source. Looking for feedback, suggestions, and potential use cases.

Hey everyone! Just a quick disclaimer for the mods: ChatGPT Microphone is a free, open-source tool, and I'm sharing it here so others can benefit from it too.

As a software engineer who chats with ChatGPT more than I talk to most actual humans (don't judge ), I wanted a more efficient way to communicate with it than having to hammer away at my keyboard all the time. I explored existing options like Voice Control and others, but their interactions felt a bit clunky and slow, with extra steps involved. They seemed more focused on hobbyists and accessibility users, rather than professionals who appreciate speed and simplicity. So, I decided to build my own Chrome extension: ChatGPT Microphone.

The extension adds a mic button to ChatGPT that looks and feels native, with CMD/CTR + M hotkeys to toggle it on/off. The most useful part IMO is “Shortcut Snippets” - when you say certain keywords, it will quick-inserts custom phrases into the conversation, like text replacement with your voice. For example, "only changes" becomes a detailed prompt asking ChatGPT for specific code changes, and "new segment" creates a paragraph break with a "---" separator.

Also, it’s free, ad-free, and open-source. Fellow devs can contribute to improving it.

I'd love to hear your feedback, suggestions, or questions. If you find it useful, let me know your use cases and what features you'd like added to the ChatGPT interface.

LINK: https://chrome.google.com/webstore/detail/chatgpt-microphone...

GITHUB: https://github.com/davidmvenegas/chatgpt-microphone

I've been looking for this! Works great thanks for sharing and making open source.

It would be cool if you could set a timeout to automatically submit the voice-to-text, e.g after 3 seconds of silence.

That's an interesting idea, I've been thinking of ways to make submitting easier (silence, kill-word, etc.) Whatever the case, it would need to be optionable ofc
Yeah option for sure, but I think it would make it feel a lot more natural UX.
Right now, I just click "CMD+M" when I'm done. But I'll add an issue to the repo and look into adding something
this is great! i didnt see a github link for it?
sry! forgot to add that, one sec...