Chat
For the cursor app, the ability to voice chat with the agent, with codebase and repo awareness, just like chatGPT voice feature.
So this tool is for planning, but with the agent already knowing about existing code and project.
When driving in the car hands free, spinning up a new idea with the agent already in the projects.
Hey, thanks for the feature request. The request is clear: a hands-free voice chat with the agent so you can plan while driving, and so the agent already knows the project.
Part of this already exists: Cursor for iOS supports voice dictation. You can dictate a prompt, it gets sent to an agent that has repo awareness via cloud agents. So you can already ask questions by voice today.
What we don’t have yet is a true conversational mode with spoken replies (TTS read-back), like ChatGPT voice. This has been requested before and it’s on our radar, but I can’t share an ETA.
Related threads if you want to add your voice or more details:
If there’s an update, I’ll follow up.
I find it totally productive to work like this. I have these long ramberling conversations in ask mode that move into plan mode with a “grill me” skill to get that shared vision nailed
@deanrie yes exactly!
My current workflow in the car, chatGPT on voice mode, initially with an instruction about the project, when I settled my idea, I have chatGPT create a .md file for cursor to start from.
I would like to skip chatGPT all together.
I’ve tinkered with a shared MCP/data store between cursor and chatGPT but it’s cumbersome.
@Dan_Cluderay yeah, that’s a common pattern: a long voice conversation in ask/plan, then a structured plan.
@Christian_Skovengaar got your workflow: remove ChatGPT from the chain and talk to the agent directly so it already knows the project. We don’t natively support that part yet: full conversational voice with spoken read-back. iOS voice dictation is already there, but a two-way voice mode is the gap we’re tracking. I can’t share an ETA yet.
Your use case and workflow description are helpful, I’ve added it to the other voice requests on this topic. If there’s an update, I’ll reply here.