Audio and video files
Attach an MP3, WAV, voice memo or video file. The audio track gets transcribed either way.
Audio to Text
Attach an audio or video file and get a timestamped transcript. Whisper v3 handles files up to 4 hours, billed per minute from 3 credits.
Audio to Text
Attach an audio or video file and get a timestamped transcript. Whisper v3 handles files up to 4 hours, billed per minute from 3 credits.
Attach an audio or video file and get a timestamped transcript back. /transcribe handles recordings up to 4 hours on Whisper v3, billed per minute from 3 credits.
Krater pricing · AI model directory · AI Image Generator and Photo Editor
Recordings in, text out
Files up to 4 hours
Whole lectures, podcast episodes and long interviews fit in one run, up to 4 hours of audio.
Timestamped segments
The transcript comes back in chunks with timestamps, so you can jump to the moment a quote happened.
Billed per minute
About 2 credits per minute of audio with a 3 credit minimum, so a short memo costs 3 and a one-hour interview about 120.
Whisper v3 underneath
Transcription runs on fal's wizper endpoint, the Whisper v3 model, measured per second of actual audio.
Transcript in the chat
The text arrives as a normal reply: copy it, translate it, or feed it into /summarize for the key points.
From recording to transcript in four steps
Attach the recording
Drop the audio or video file into the chat. Anything from a voice memo to a 4-hour lecture works.
Whisper v3 transcribes it
The audio runs through the wizper endpoint and comes back as timestamped text.
Read the transcript
Segments arrive with timestamps so you can find the exact moment something was said.
Pay per minute of audio
About 2 credits per minute, 3 credit minimum. The estimate shows before it runs.
Who transcribes audio
Interviews
Turn recorded conversations into quotable, citable text.
Lectures
Get study notes from a recorded class instead of re-listening.
Podcasts
Make episodes searchable, quotable and easy to clip.
Meetings
Convert recordings into minutes, decisions and action items.
Voice memos
Turn quick dictations into text you can file and search.
Copy a prompt, run it as /transcribe
When a different tool fits better
Noise and crosstalk degrade it
Clean speech transcribes accurately; heavy background noise, music and people talking over each other introduce errors. Check quotes before publishing.
No speaker labels
The transcript is one continuous voice track with timestamps, not labeled turns. For who-said-what diarization, a dedicated meeting tool fits better.
One-off memos stay cheap
There is no free plan for transcription. Pro starts at $20 per month. For a 30-second memo the 3 credit minimum is the whole cost.
Audio to text questions
Is transcription free to use?
Not fully. /transcribe bills per minute of audio, about 2 credits per minute with a 3 credit minimum. Pro is $20 per month and includes monthly credits.
What files can I transcribe?
Audio files (MP3, WAV, voice memos) and video files both work; the audio track gets transcribed. Recordings up to 4 hours fit in one run.
Does the transcript have timestamps?
Yes. The text comes back in segments with timestamps so you can locate the moment any quote happened.
Can it tell different speakers apart?
Not today. You get one continuous timestamped transcript, not speaker-labeled turns. For diarization, a meeting-specific tool fits better.
Related tools and models
- Text to Speech
Turn written text back into natural audio in 32 languages.
- AI Summarizer
Condense the transcript into the key points.
- AI Document Editor
Format the transcript into a clean document.
- AI model directory
Every model on Krater with credit pricing and capabilities.