Voice and text now live in one place instead of two
A room is a running conversation: people talk, people type, and the record of it should read like one continuous thread. Until now, transcripts and text messages felt like separate streams. We've rebuilt the activity panel from the ground up so they aren't.
Everything lives in one timeline. Voice transcripts, typed messages, and recording-only notes flow in chronological order, grouped by speaker into clean sessions with day dividers. A small icon tells you at a glance which is which: a mic for voice, a keyboard for text, and your own messages carry a subtle Signal Orange tint so the conversation is easy to follow.
Voice is finally first-class. Every spoken line carries a mini-waveform and a play button, so you can hear the exact audio without leaving the stream, or open the full replay to scrub, skip, and adjust volume. While a teammate is talking, their words stream in live with a pulsing indicator, then finalise the moment they release push-to-talk. Confidence and language cues surface only when they matter.
The timeline is built to read fast. Search every word in a room and jump straight to playback. Switch between Expanded bubbles for review and Compact rows for scanning. Square corners, flat inset rings, and JetBrains Mono on every timestamp: the redesign is unmistakably whoot., right down to the pixel.
Designed to be read at a glance
One stream carries every message type, with replay, live partials, and search built in.
- One unified timeline
- Voice transcripts, text messages, and recording notes in one stream
- Grouped by speaker into sessions with clear day dividers
- Auto-scrolls to the live moment: scroll up to review without losing your place
- Voice, made legible
- A mic, keyboard, or muted-mic icon tells voice from text from recording-only
- Confidence dots surface only when a transcript dips below 85%
- Language badges appear automatically for anything that isn't English
- Replay inline
- Every voice message carries a mini-waveform and a play button
- Open the full replay for scrubbing, ±5s skips, and a volume slider
- Speaker, room, and timestamp stay pinned to the audio
- Transcribed live
- Partial transcripts stream in real time while a teammate is talking
- A pulsing emerald LIVE indicator marks the active utterance
- The transcript finalises the instant push-to-talk is released
- Search every word
- Search across a room's messages and transcripts in one popover
- Matches are highlighted, with speaker and timestamp on every result
- Jump straight to playback from any voice result
- Expanded or compact
- Expanded bubbles for reading and review: full metadata, grouped by speaker
- Compact rows for fast scanning: one dense line per utterance
- A single toggle switches density without losing your position
Read, play back and search every message in one timeline.
Designed to match the rest of whoot.
The redesign follows the whoot. design system to the pixel. Corners are square everywhere, flat inset rings replace drop shadows, and every value is set in JetBrains Mono. Signal Orange is held back for the things that matter: your own messages and the live controls. The result is a stream that's dense where it needs to be, quiet where it doesn't, and instantly scannable throughout.
- Square corners everywhere: zero border-radius, true to the whoot. design system
- Ring-not-shadow elevation: flat inset rings instead of soft drop shadows
- JetBrains Mono for every value: timestamps, durations, confidence, and language codes
- Signal Orange is reserved for your own messages and live controls, never decoration
- Speaker avatar chips, session headers, and day dividers make the stream scannable
- Confidence and language cues appear only when they matter, never as clutter
- A branded empty state: a waveform mark and a clear prompt, not a blank box