Greasy Fork

来自缓存

Improved Claude Voice Interface

Voice interface for Claude with TTS functionality, speech recognition, auto-reading and drag positioning

作者
Stephen Tremaine
日安装量
0
总安装量
2
评分
0 0 0
版本
0.7
创建于
2025-02-27
更新于
2025-03-04
大小
28.5 KB
许可证
MIT
适用于

This is still pretty rough but it's my attempt to enable something like the conversation mode of chat gpt in claude.

Assuming they haven't changed the website and broken my references, it should allow you to hit the listen button at which point it will enable the browser's speech to text engine continuously updating the text/chat box until you hit the listen button again, at which point it will try and submit your message. You should be able to edit the text as you go. It will remove white space/line breaks as it updates though and depending how quickly you hit the stop recording button, the submit button might not be available to press yet. Just click it yourself if this happens. It adds a period and space between statements that it hears. That's not always ideal if you're like me and tend to pause to think for a while but it doesn't seem to hamper claude's understanding too much.

For new messages, it will try and detect them and read them out load using the browser text to speech engine.

If you click and drag the icon on the far right, it will move the interface up and down out of your way if it's on top something you want to see/click/etc.

Note - you will need to give microphone permissions for this to work.

This is far from my best work so I'm open to feedback for improvement - I started this mostly to test claude's programming ability and was actually writing it from my phone. It generated some very messy code that I've tried to refactor but it's still little messy. I'm mostly sharing so I can keep it in sync between my desktop and phone but feel free to try it out and give me some feedback if you want to help me unwind the messy code!