Transcription · How-to
How to transcribe audio to text with speaker labels and SRT captions
Transcribe interviews, podcasts, and videos into text with speaker labels, rename speakers, and export TXT, SRT, or VTT captions. Tips for better accuracy.
When speaker labels matter
For interviews, meetings, and podcasts, a wall of text is hard to use. Speaker labels split the transcript by voice ("Speaker 1", "Speaker 2"), so you can quote people correctly, write show notes, and create captions that make sense.
Step 1: Upload your recording
VoiceLamp accepts WAV, MP3, M4A, MP4, MOV, FLAC, OGG, Opus, and WebM. Files can be up to 5 minutes and 25 MB. For video, the audio track is used automatically.
Choose the spoken language yourself when you know it. For Urdu or Arabic, always select the language explicitly. If speakers switch languages, choose Mixed language.
Step 2: Turn on speaker identification
Turn on Identify speakers before you get the price. Each part of the transcript is labelled with the speaker who said it.
Step 3: Review and rename speakers
Open Review transcript to read through the text, rename "Speaker 1" to a real name, and mark the transcript as reviewed. Your names are used in every download.
Step 4: Export TXT, SRT, or VTT
- TXT for articles, show notes, and quotes.
- SRT for YouTube, Facebook, and most video editors.
- VTT for web video players.
Tips for better accuracy
- Record close to the microphone and avoid background music.
- Clean up noisy recordings first with AI voice cleanup.
- Split long recordings into parts of up to 5 minutes.
What it costs
Transcription costs about 4,400–5,600 credits per minute, and you see the exact price before it runs. If the main engine has a problem, the job moves to a backup engine automatically at no extra cost.
Try it yourself. See the exact price before anything runs, and get credits back automatically if a job fails.
Open transcription