Audio cleanup, transcription, emotion tags, and data privacy.
How does AI audio cleanup work?
Our AI removes background noise, echo, and crowd hum, and evens out the volume. Cleaning usually takes a few minutes.
Can I preview before/after cleanup?
Yes, both the original and cleaned versions are available. Play either one from the track list.
Can I revert a cleaned track?
Yes, the original file is always preserved. You can re-clean or download the raw original at any time.
How does AI transcription work?
We use AI to convert speech to text for every audio message. Transcriptions appear below each track and are included in PDF exports.
What languages does transcription support?
Transcription supports 15 languages: English, Spanish, Portuguese, French, German, Italian, Japanese, Korean, Chinese (Mandarin), Arabic, Hindi, Russian, Polish, Dutch, and Turkish, plus Auto-detect. Auto-detect (the default) is accurate for most events. Set a specific language when you know your guests will primarily speak one language, it improves accuracy and means guest name recognition runs in that language too.
What are emotion tags?
AI analyzes the transcribed text to detect the emotional tone: Heartfelt, Funny, Celebratory, or Nostalgic. Tags appear as badges on each track.
How accurate are emotion tags?
Emotion tags are based on the words spoken, not vocal tone. They’re generally accurate but should be considered suggestions, not definitive labels.
Is my audio data private?
Yes. Your audio is processed ephemerally and is not used to train any AI models. Files are stored securely in your private storage bucket.