Feature request for product/service
Cursor Web
Describe the request
Add native voice input and voice output for supervising, reviewing, and directing agent work.
As agents become more capable, the bottleneck is increasingly not generating work but reviewing work. I often have multiple agents running in parallel and would like to be able to monitor progress, review results, make decisions, and redirect work while away from my desk.
The ideal experience is putting on a Bluetooth headset, going for a walk, and interacting with my active agent tasks entirely through voice.
Example workflows:
“Give me a status update on all active agents.”
“What changed since my last review?”
“Summarize the work completed by Agent 2.”
“Read me the implementation plan.”
“Compare the outputs from the three agents working on this feature.”
“Which solution do you recommend?”
“Continue with option B.”
“Ask that agent to investigate performance implications.”
“Create another agent to evaluate an alternative approach.”
“Read me the pull request summary.”
Voice output is just as important as voice input. The goal is not dictation. The goal is conversational review and supervision of agent work.
Potential capabilities:
Natural voice conversations about active tasks
High-quality text-to-speech responses
Bluetooth headset support
Mobile support through Cursor Web
Agent status briefings
Voice review of reports, plans, and completed work
Side-by-side verbal comparison of agent outputs
Voice approval and redirection of tasks
Management of multiple agents in parallel
Ability to start work on desktop and continue supervision from mobile
As agents become increasingly autonomous, developers spend more time reviewing and directing work than producing it themselves. Voice-first supervision would allow users to remain productive while walking, commuting, exercising, or otherwise away from the keyboard.
The dream workflow is simple:
“I want to go for a coding walk with a Bluetooth headset and review, manage, and direct my team of agents.”