Continuous voice mode!

Ideas Feature Requests
1

Feature request for product/service

Cursor Web

Describe the request

Current problem: I have to click the voice icon, then a send icon, each time I want to dictate a message to cursor web.

Current workflow:

  1. Click voice icon
  2. Speak
  3. Click finished icon
  4. Click send message icon

Ideal solution: I click the voice icon once, there is an additional button for continuous voice mode I can press and then it will send messages once I have stopped talking for a good 3 seconds or so, or have reached the end of a sentence, and send that as the input. It will stay in this mode until I leave it, or it has been 60 minutes to turn it off from listening automatically. This would make the experience MUCH nicer, it is how Codex’s mobile app works. I have android so can’t use the iOS app and have no idea if the iOS app works that way already.

Ideal workflow:
[once already in continuous mode]

  1. Speak
5

Hey, thanks for the detailed feature request. You laid it out step by step, which is super helpful.

Right now in Cursor Web, voice works like dictation: record → transcript gets inserted into the input → you send it manually. There’s no continuous mode with auto-send on silence (VAD) on any surface yet. So your ask is valid.

On iOS, voice is also dictation-only for now, so there’s no always-on mode like you described on mobile either.

Voice-first workflows have come up before. There are a couple related threads discussing similar scenarios. If you want, add your thoughts there too, it helps with prioritization:

No timeline to share yet, but if there’s an update I’ll post it in the thread.