Transcription
Transcription is the foundation for nearly every follow-up workflow in Hyper. It creates the line-by-line meeting record that powers captions, summaries, action items, and search.
Use transcription when the team needs accurate recall, speaker attribution, and a reliable source of truth after the call ends.
Wireframes
Section titled “Wireframes”┌──────────────────────────────────────────────────────────────────────┐
│ Hyper app.hyper.video/room/launch-review │
├──────────────────────────────────────────────────────────────────────┤
│ │
│ Meeting room controls │
│ │
│ [ Mic ] [ Camera ] [ Share ] [ Chat ] [ Participants ] │
│ [ Notes ] [ Captions ] [ Record ] [ More ] │
│ │
│ Stage │
│ ┌──────────────────────────────────────────────────────────────┐ │
│ │ Presenter: Maya Chen │ │
│ │ Shared screen: Launch checklist and open blockers │ │
│ └──────────────────────────────────────────────────────────────┘ │
│ │
│ Side panels │
│ - Participants waiting: 2 │
│ - Raised hands: 1 │
│ - Summary capture: active │
│ │
└──────────────────────────────────────────────────────────────────────┘
Captions and transcript controls are visible during the meeting so participants can follow live instead of waiting for the recap.
When to use this
- Meetings where exact wording, decisions, or commitments matter later.
- Sessions that need searchable follow-up without replaying the entire recording.
- Cross-team calls where not every participant can attend live.
Before you start
- Confirm participants understand whether transcription is on by default.
- Choose the artifact permissions and retention policy that fit the meeting type.
- Check microphone quality if you expect noisy environments or many remote participants.
Settings and options
Transcript behavior
- Use captions for live readability and the transcript panel for detailed review.
- Treat speaker attribution as a strong starting point and edit lines when exact labels matter.
- If translated transcripts are enabled, verify the target language before the meeting begins.
What this feature helps with
Speaker identification
Each speaker is automatically labeled as they talk. No manual tagging — see who said what at a glance, with timestamps for every line.
Searchable timelines
Jump to any moment by searching the transcript. Find decisions, commitments, and context across your meeting history in seconds.
Feeds other AI features
Transcripts power summaries, action items, and search. One source of truth — your meeting content flows into every AI feature automatically.
How to work with it
Enable transcript
- Transcription is on by default when you join a meeting
- No setup or configuration required — it starts as soon as you speak
Live speaker-labeled transcript appears
- Watch the transcript build in real time as participants talk
- Each line is tagged with the speaker and timestamp
Jump to timestamps
- Click any line to jump to that moment in the recording
- Use search to find specific moments across the transcript
Transcript feeds summaries + search
- Summaries and action items are generated from the transcript
- Full-text search indexes every word for later discovery
Where it fits well
Team meetings
- Capture every discussion with speaker attribution
- Search past meetings for decisions and context
- Share transcripts with attendees who couldn't join
- Use as input for summaries and action items
Client calls
- Professional transcripts for follow-up and documentation
- Find commitments and next steps with speaker labels
- Keep sensitive conversations private — all on-device
- Export for records or sharing with stakeholders
Training sessions
- Full transcripts for onboarding and training materials
- Searchable reference for learners to revisit key points
- Speaker-labeled so instructors can review their delivery
- Feeds into summaries for quick recap distribution
What to expect
- On-device processing
- 97% accuracy
- Speaker diarization
- Real-time with <100ms delay
- Searchable after meeting
- Powers summaries and search
Trust and controls
- On-device only
- Encrypted storage
- Per-meeting toggle
- Participant notification
Questions that come up often
How accurate is the transcription?
Hyper achieves 97% accuracy for clear speech in supported languages. Accuracy may vary with background noise, accents, or technical jargon. You can always review and edit transcripts before sharing.
How does it handle accents and dialects?
Our on-device models are trained on diverse accents and dialects. Accuracy is generally strong across regional variations, though highly specialized accents may see slight variance. You can edit any transcript line if needed.
Can I disable transcription?
Yes. You can turn off transcription per meeting or account-wide. Toggle it off before or during a call. When disabled, no transcript is generated and no audio is processed for transcription.
Who can see the transcript?
By default, transcripts are visible to meeting participants. You control sharing: keep them private, share with attendees only, or distribute more broadly. Workspace admins can configure default visibility.
What export formats are available?
You can export transcripts as plain text, SRT (for subtitles), or structured formats for integration with other tools. Export is available from the meeting recap view after the call ends.
Related tasks
Limitations and rollout notes
- Accuracy depends on microphone quality, overlap, and language support.
- Search and summary workflows depend on transcript availability and artifact permissions.
Troubleshooting
Transcript quality is lower than expected
- Check whether the affected participants were using the correct microphone.
- Reduce overlapping speech and heavy background noise where possible.
- Review whether translation or a different language setting should have been enabled.