Skip to main content

Automatic Voice Note Transcription with AI

Updated over 2 weeks ago

The Tractian platform converts voice recordings into text using transcription-based artificial intelligence. This streamlines information logging and makes it easier for your entire team to access. Transcription quality depends directly on the clarity of the recorded audio.


Where the transcription appears

Every time you record a voice note on the platform, the system automatically starts the transcription process, and the resulting text is displayed just below the audio player within the same record.


While processing, the message “Transcribing audio…” indicates that the file is being analyzed.


Once the transcription is complete, the text appears below the audio along with the question “Was this transcription helpful?” for user feedback.


Available controls

Within the audio player and transcription section, you’ll find:

  • Playback speed options: 1x, 1.5x, and 2x

  • Text actions: Copy transcription and Download audio (in mp3 format)


Transcription feedback

After each transcription, you can rate it as Helpful or Not helpful.


If you select “Not helpful,” the system will prompt you to specify the reason, such as:

  • Missing words

  • Incorrect words

  • Missing punctuation

  • Incorrectly spelled names

  • Transcription delay

  • Wrong language

  • Other reason

This feedback helps us continuously improve the transcription AI.


Best practices

To ensure the highest transcription quality:

  • Record in quiet, echo-free environments

  • Speak clearly and at a steady pace, close to the microphone

  • Avoid multiple people speaking at once


What to do if transcription fails

During or after processing, you may see one of the following messages:

  • “Failed to transcribe audio” — an error occurred; try again

  • “Transcription unavailable” — the transcription could not be completed

If this happens:

  • Play the audio to confirm it’s audible

  • Click “Try again” (if available) to reprocess the file

  • If the issue persists, re-record the audio in better conditions (less noise and closer voice input)

Did this answer your question?