Tengos converts spoken audio to text in real time using advanced speech recognition. Transcripts appear on screen as speakers talk, giving you an instant, readable record of everything that's said.
How live transcription works
Transcription happens in two stages:
- Interim results — a provisional version of each sentence appears as the speaker talks. These may shift or correct themselves as Tengos processes more of the audio.
- Final segments — once Tengos confirms a segment, it locks in the final text. Confirmed segments are stable and will not change.
This two-stage approach allows Tengos to display text with minimal delay while maintaining high accuracy in the final transcript.
During transcription
- Transcript segments appear continuously in the main panel as the meeting progresses.
- Each segment is attributed to the correct speaker in Multi-Speaker sessions.
- You can scroll back through the transcript at any time without interrupting the live feed.
- Auto-scroll keeps the view at the latest segment when enabled.
Supported languages
Tengos supports transcription in 85+ languages. Select your source language before starting the meeting, or use Auto-detect to let Tengos identify the language automatically.
Tips
- Speak clearly and at a steady pace for the best accuracy.
- Use a dedicated external microphone rather than a built-in laptop mic where possible.
- Reduce background noise — closed rooms with minimal ambient sound produce significantly better results.
- Create a Glossary with domain-specific terms (brand names, technical jargon, product names) to improve recognition accuracy for specialized vocabulary.
Notes
- Tengos automatically handles brief internet disconnections. Transcription pauses during an outage and resumes automatically when the connection is restored.
- After the meeting ends, you can search, copy, and export the full transcript from the meeting detail view.