[FEATURE] Two-way voice mode in Cowork (not just dictation)

Status Open
Maintainer reply None cached
Activity 1 comment · opened Oct 2, 2026

Preflight Checklist

  • [x] I have searched existing requests — related: #76097 and #90287 (see Additional Context for why I'm filing this separately)
  • [x] This is a single feature request (not multiple features)

Problem Statement

I'd love to have two-way voice mode (talk and hear Claude reply) in Cowork sessions, not just dictation.

Right now Cowork only offers dictation, so Claude's replies are text-only. Voice mode is limited to regular Chat, which means I have to leave my Cowork workspace whenever I want a spoken conversation — and lose the context, files and scheduled tasks that live there.

Proposed Solution

Bring the same two-way voice mode that exists in regular Chat to Cowork sessions: I speak, Claude replies out loud, within the same Cowork conversation and with the same workspace context (files, connectors, tools, scheduled tasks) available.

Alternative Solutions

  • Dictation in Cowork + reading replies on screen: works, but isn't hands-free and isn't a conversation.
  • Switching to regular Chat for voice: loses all the Cowork context, so I have to re-explain everything.

Priority

Medium - Would be very helpful

Feature Category

Other

Use Case Example

  • Talking through a plan while walking or commuting.
  • Reviewing results hands-free.
  • Brainstorming ideas out loud.
  • Practicing a presentation or interview and getting spoken follow-up questions — with Claude having my materials from the Cowork workspace at hand.

Additional Context

Environment: Claude Desktop on macOS. In Cowork I only see dictation; there is no two-way voice mode.

Related issues: #76097 (continuous voice conversation in Desktop/Cowork) and #90287 (full two-way voice for Cowork) ask for the same capability. I'm filing this to add the use cases above and the macOS data point. #97954 suggests voice mode may already be appearing in Cowork on Windows — if it's rolling out, it would be great to have it on macOS too, with tools/connectors available while in voice mode.

Thanks for considering it!

View original on GitHub ↗

This issue has 1 comment on GitHub. Read the full discussion on GitHub ↗