How to Transcribe
The basic flow with screenshots: upload a recording, get speaker-separated transcription, edit, export, and generate voice.
Supported Formats & Limits
Formats: MP3 / WAV / M4A / WEBM / FLAC / MP4
Per file: up to 2 hours and 500MB
Free plan: 60 transcription minutes + 10,000 TTS characters monthly (no credit card). Try up to 90 seconds on the top page before signing up.
1.Upload an audio file
Drag & drop onto "New transcription" in the workspace (/transcripts), or choose a file. If you have no audio at hand, "Try with sample audio" walks the same flow.
2.Transcription runs automatically
Processing starts automatically after upload, and the list status (queued → processing → done) updates itself. When finished, the result is stored as timestamped segments with speaker labels.
3.Edit in the editor
Open the detail view to edit with an audio-synced editor. Click a segment to play from that point; renaming a speaker applies to the whole transcript. In-document search helps find misrecognitions quickly.
4.Export (TXT / SRT / VTT)
Export speaker-labeled TXT for minutes, or SRT/VTT for video subtitles. Clipboard copy pastes straight into minutes or chat.
5.Turn text into voice (Speech)
Paste text into Speech (/speech) to generate natural narration (MP3). Edited transcripts can be voiced as-is. Higher plans unlock emotion tags and more voices.
Tips for better recordings (5 points to improve accuracy)
- Get mics close to speakers — one phone picking up a whole room hurts speaker separation
- Reduce crosstalk — just having a facilitator ask for one-at-a-time speaking saves editing time
- Record a 30-second test and check the quietest person is audible
- Record in standard formats (MP3/M4A); convert unusual codecs before uploading
- Get consent to record — telling participants about the recording and its purpose is the professional standard
About your data
- Audio and transcripts are stored only under your account and never shared with other users
- They are never used as AI training data
- Delete your data any time from the workspace
Related
Try transcription for free
The Free plan includes 60 minutes of transcription and 10,000 characters of voice generation every month. No credit card required.
Sign up free