Drop your audio file into fwip’s Speech to Text tool. On-device AI transcribes it — speaker labels, timestamps, the lot. Your recording never leaves your machine. No upload. No server. No third party listening to your meeting. Download the transcript as text or copy it straight out.
How to do it
- Open fwip’s Speech to Text tool.
- Drop your audio file in — MP3, WAV, M4A, OGG, FLAC, WebM, or MP4.
- Select the language if it’s not English.
- Hit Transcribe.
- Read, edit, copy, or download the transcript.
The AI model runs in your browser. Your audio file never touches a server.
Why this matters
Every major transcription service — Otter, Descript, Rev, HappyScribe — uploads your audio to their servers for processing. That’s fine for a podcast episode. It’s not fine for a recorded legal consultation, a sensitive HR meeting, a medical appointment, a board discussion, or a therapy session.
The question isn’t whether these companies are trustworthy. The question is whether the recording should leave your device at all. For some recordings, the answer is no.
fwip’s transcription runs entirely on-device using WebAssembly-compiled AI models. The audio stays on your machine. The transcript is generated locally. Nothing is transmitted.
Frequently asked questions
Is the transcription really happening on my device? Yes. The AI model loads into your browser via WebAssembly. Processing happens using your device’s CPU. No network request is made with your audio data. You can verify this by disconnecting from the internet after the page loads — it still works.
How accurate is on-device transcription? For clear audio with one or two speakers, 90–95% accuracy. Noisy recordings, heavy accents, or overlapping speakers reduce accuracy. It’s comparable to Otter’s free tier but without the upload.
Can it identify different speakers? Yes. fwip’s Speech to Text includes speaker diarisation — it labels Speaker 1, Speaker 2, etc. based on voice differences.
What’s the maximum file length? Depends on your device’s memory. Most modern laptops handle 60–90 minutes comfortably. For longer recordings, the desktop app is more reliable.
What languages are supported? English and major European languages. Check the tool page for the current list.
How is this different from Otter or Descript? Otter and Descript upload your audio to cloud servers for processing. fwip processes everything on your device. Your recording never leaves your machine. That’s the difference.