[FEATURE] ature request: voice (text-to-speech) output in Cowork and Claude Code

Preflight Checklist

• [x] I have searched existing requests and this feature hasn't been requested yet
• [x] This is a single feature request (not multiple features)

Problem Statement

Voice output (text to speech) exists in the Claude chat app but not in Cowork or Claude Code. I do most of my work in those two tools, including a captioning app I built for myself in Claude Code. What I can't do there is hear text read back while I work. To listen to something I have to copy it out into a separate tool, which breaks the workflow every time. I also have cerebral palsy and I'm an auditory learner, so hearing text is how I catch what's working and what isn't.

Proposed Solution

Bring the voice output that already ships in chat to Cowork and Claude Code, with one critical difference from how chat voice works today: voice output needs to be independent of a hands-free voice mode. In chat right now, turning voice on drops you into a voice conversation that disables typing. I need the opposite, the ability to keep typing or dictating my side while Claude's responses are read aloud automatically. Make it on-demand and selectable too, so in Code I can have prose and explanations read back but never the source code itself.

Alternative Solutions

_No response_

Priority

Medium - Would be very helpful

Feature Category

Interactive mode (TUI)

Use Case Example

Here is exactly how I would use it, and it's different from chat's current voice mode.
I type using Mac's built-in dictation, not in-app voice input, because Mac handles my cerebral palsy speech far better and gives a much cleaner transcription than the system trying to interpret my voice. What I want from Claude is the output side: responses read back to me automatically while I keep dictating and typing.
I already do this with Grok. I leave its voice output on, mute my mic, in the grock interface and dictate my input through Mac. I get automatic playback of every response without doing anything extra, and the input stays clean because Mac is doing the transcription.
The problem with chat voice today is that turning it on puts you in a voice mode that disables typing. It becomes a hands-free conversation instead of "type normally and hear the replies," which is much less useful for real work. So the ask is voice output as a playback layer that coexists with text input, in Cowork and Code.

Additional Context

I'm disabled and I build things with these tools, and I think a lot about technology and disability from lived experience. Glad to share a user's perspective if it's ever useful.