Transcribe with Distil-Whisper
5.8%
WER
99
Languages
48.0x
Speed
MIT
License
About Distil-Whisper
Distil-Whisper is a distilled version of Whisper created by Hugging Face. It reduces the model size by 49% and achieves 6x faster inference while maintaining within 1% WER of the original Whisper Large V2 on out-of-distribution evaluation sets.
Model Info
- ProviderHugging Face
- Architecture-
- LicenseMIT
- UpdatedMar 2026
Frequently Asked Questions
Distil-Whisper is a speech-to-text model by Hugging Face. STT.ai hosts Distil-Whisper on our GPU infrastructure so you can use it without provisioning your own hardware — upload audio or video and pick Distil-Whisper from the model picker.
On standard benchmarks, Distil-Whisper achieves around 5.8% Word Error Rate. Real-world accuracy depends on audio quality, accent, and language; for noisy or accented recordings, expect a few percentage points higher WER.
Distil-Whisper runs on STT.ai's free tier — every visitor gets 600 minutes/month at no cost. Paid plans add longer per-file limits, private transcripts, and priority queueing.
Distil-Whisper is released under MIT, a permissive open-source license. You can self-host Distil-Whisper on your own hardware or use our hosted version — both are commercially usable.
Distil-Whisper supports 99 languages. Auto-detection picks the right language for most audio; you can also specify it manually for a small accuracy lift.
Distil-Whisper processes audio at about 48.0x real-time on our GPUs. A 1-hour audio file finishes in under 1 minutes; longer files queue and notify by email when done.
Distil-Whisper has 756M parameters. Larger models tend to be more accurate but slower; STT.ai hosts Distil-Whisper on GPU so the parameter count doesn't affect your client-side performance.
Distil-Whisper accepts every format STT.ai supports — MP3, WAV, M4A, FLAC, OGG, MP4, MKV, MOV, WebM, AVI, and others. Output as TXT, SRT, VTT, DOCX, JSON, or PDF.
Yes. Speaker diarization runs alongside Distil-Whisper for every transcription — each speaker is labeled and you can rename them in the editor afterwards.
Yes. Distil-Whisper runs in our managed environment — audio is processed and deleted by default and never used for training without explicit opt-in. Pro plans add client-side encryption for transcripts at rest.
Use the compare-stt tool to run Distil-Whisper against any other supported model on the same audio — you'll see WER, segment count, speaker labels, and confidence scores side-by-side. The Distil-Whisper vs Whisper Large V3 comparison is the most commonly run.
Yes. Specify "distil-whisper" as the model parameter on the /v1/transcribe endpoint. Python and Node.js SDKs include Distil-Whisper examples. Free API tier includes 100 minutes/month.
Yes. Because Distil-Whisper is MIT-licensed, you can self-host it. STT.ai's open-source page lists the project repo and weights. Most production teams use our hosted version to skip GPU procurement, model swaps, and ops.