Changelog/Voice Mode — Create Cards & Polls by Speaking
v2.9.0featureai

Voice Mode — Create Cards & Polls by Speaking

Retro notes usually happen in two places: on the board, and in the room while everyone is talking. Starting today, Voice Mode bridges the gap — just speak and letRetro does the rest. Say what's on your mind, create cards, spin up polls with deadlines and priorities, all hands-free. No typing required.

How it works

  • Dictate, don't type — Hold Shift + Space in any board and speak. LetRetro transcribes your words and turns them into a card
  • Create cards by voice — Say "add a card: onboarding felt confusing, priority high" and letRetro creates it with the right priority, no clicks needed
  • Create polls with deadlines & priority — Say "create a poll: should we extend the sprint? deadline Friday 5pm, priority urgent" — letRetro drafts the poll, sets the deadline and priority automatically
  • Deadlines, priorities, assignees — just say it — Mention a due date, priority level, or assignee while you speak and AI parses it into structured fields
  • Smart column placement — AI reads the board and suggests the most fitting column for each card, so "onboarding felt confusing" lands in What went wrong instead of sitting in the wrong pile
  • Auto mode — Skip the confirmation dialog entirely. Cards and polls are added directly to the suggested place, with a quick visual confirmation when they land
  • Manual picker — Prefer to choose? Review the transcript, edit it, and pick the column, deadline, or priority yourself before adding
  • Up to 2 minutes per card — Long enough for a full thought, short enough to keep things moving
  • Understands context — The AI reads your board's columns and existing cards before suggesting placement, so suggestions fit your retro's structure

How to enable it

Voice AI is available on all plans — Free includes 10 voice cards per month, Pro includes 100:

  1. Open your board and press Shift + Space to start recording
  2. Speak your note and release to stop
  3. Review the transcript, or turn on Auto mode to add directly to the suggested column

You can manage Auto mode from the voice widget or under Board Settings → Features.

Audio is transcribed with OpenAI's Whisper and never used for training. Cards are sent to the model only to find the right column.

What's next

We're working on more languages for transcription and voice commands (like "start a poll" or "create a document") so you can run more of the board hands-free.