Transcribe with Distil-Whisper

Name: Distil-Whisper
Author: Hugging Face

Àwọn iṣẹ́ láti mú àwọn àwòrán àti àwòrán tí a yàn fún gbogbo eniyan. Àwọn àwọn ìròyìn tí a dáwọ́ láti lo DRM kò fọwọ́sì.

Àwọn ìṣàfihàn fún àwọn ìṣàfihàn

Private transcript

Fi àkọlé pamọ́

Ṣí àwọn àwọn àgbéwọlé →

Tí fáìlì náà síbẹ̀ tàbí tẹ̀ láti ṣàfihàn

MP3, WAV, M4A, FLAC, MP4, MKV, MOV, WebM — to 2GB

Fi àwọn fáìlì mìíràn pamọ́ Àwọn àwọn àwọn àwọn àwọn

Àwọn ìṣàfihàn fún àwọn ìṣàfihàn

Private transcript

Fi àkọlé pamọ́

Ṣí àwọn àwọn àgbéwọlé →

Àwọn ìṣàfihàn fún àwọn ìṣàfihàn

Àwọn àkọlé àìpẹ́ láti inú àkọlé. AI kọ̀ọ̀kan-ìṣàmúlò-ètò bí a tí n sọ̀rọ̀ - ìṣàmúlò-ètò náà tí a bá kọ̀ọ̀kan àwọn àkọlé náà.

Àwọn àwọn àmì-ìwé

10 free min/day 600 min ọfẹ pẹlu iforukọsilẹ Kò ní kaadí ẹ̀yàn Àwọn àmì-ìwé

Ṣẹ̀dà nípa ọ̀fẹ̀ →

5.8%

WER

Languages

48.0x

Speed

MIT

License

About Distil-Whisper

Languages Supported by Distil-Whisper

English

Spanish

French

German

Chinese (Mandarin)

Japanese

Korean

Portuguese

Arabic

Hindi

Russian

Italian

Dutch

Turkish

Polish

Swedish

Indonesian

Thai

Vietnamese

Czech

Greek

Romanian

Hungarian

Hebrew

Danish

Finnish

Norwegian

Ukrainian

Malay

Bengali

Model Info

ProviderHugging Face
Architecture-
LicenseMIT
UpdatedMar 2026

Related Models

3.2% WER

4.2% WER

5.1% WER

3.5% WER

7.8% WER

Àwọn Àtòjọ-ẹ̀yàn

Distil-Whisper ní móòdù ìṣàfihàn-si-àkọ́kọ́ láti Hugging Face. STT.ai ní àwọn awáròyìn Distil-Whisper nípa àgbègbè GPU wà láti jẹ́ pé o lè lò láti fi àwọn àwọn ìṣàfihàn rẹ̀ pamọ́ - fi àwòrán àti àwòrán pamọ́ sínú àwọn awáròyìn Distil-Whisper láti inú àwọn àwọn awáròyìn móòdù.

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 nlọ lori STT.ai ti o ni ọfẹ — gbogbo awọn alejo gba iṣẹju 600/oṣu kan laisi idiyele. Awọn iṣẹju-aaya ti a sanwo ṣafikun awọn ibiti o ju-faili, awọn atẹwe ti ara ẹni, ati awọn atẹwe ti o ni ifẹ.

Distil-Whisper tí a fi pamọ́ láti inú MIT, ìlàyè ìṣàfilọ́lẹ̀-ìṣílọ́lẹ̀. O lè fi Distil-Whisper pamọ́ sípàrà rẹ̀ láti lò nínú àwọn ìṣàfilọ́lẹ̀ wà - gbogbo wọn ní a lè lò nínú iṣẹ́.

Distil-Whisper ǹfà àwọn ìtàn 99. Àwọn ìṣàfihàn-ìdáràn àwọn ìtàn tí a fẹ́ fún àwọn ìṣàfihàn àwọn ìranlọwọ; o lè sọ̀rọ̀ nípa ìrànwọ́ fún ìṣàfihàn àwọn ìṣàfihàn àwọn ìṣàfihàn.

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 gba gbogbo àwọn ìrísí-lẹ́tà tí STT.ai ǹfàyè — MP3, WAV, M4A, FLAC, OGG, MP4, MKV, MOV, WebM, AVI, àti àwọn mìíràn. Àwọn ìṣàfilọ́lẹ̀ bí TXT, SRT, VTT, DOCX, JSON, tàbí PDF.

Ya. Ìṣàfilọ́lẹ̀ àwọn àkọlé àwọn àkọlé náà tí wọ́n lọ́wọ́lọ́wọ́ Distil-Whisper fún ìṣàfilọ́lẹ̀ gbogbó - àwọn àkọlé àwòrán ní pàtó àwọn àkọlé àwòrán náà nínú àwọn àwọn àkọlé àwòrán.

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.

Ya. Ṣàfihàn "distil-whisper" bí àwọn ààtò ìṣàmúlò-ètò módè́èlì lórí àwọn ààtò ìparí iṣẹ́ /v1/transscribe. Python àti Node.js SDKs ní àwọn ààtò ìṣàmúlò-ètò Distil-Whisper. Àwọn ààtò API àìfẹ́ ní àwọn ààtò 100 àwọn àkókò/óṣù.

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.

Transcribe with Distil-Whisper

About Distil-Whisper

Languages Supported by Distil-Whisper

Model Info

Related Models

Àwọn Àtòjọ-ẹ̀yàn

Kini Distil-Whisper?

Bawo ni Distil-Whisper ṣe tọ́?

Njẹ Distil-Whisper jẹ ọfẹ lati lo?

Ààyè wo ni Distil-Whisper lò?

Ìgbà wo nínú àwọn ìtàn wo ní Distil-Whisper ǹfà?

Ìgbà wo ní Distil-Whisper jẹ́?

Igbà wo ní módè́lì Distil-Whisper náà jẹ́?

Kini àwọn ìrísí-lẹ́tà àwọn orin lè kọ̀ nípa Distil-Whisper?

Ń Distil-Whisper gbọ́ àwọn ìgbàkọ̀lù àwọn ìgbàkọ̀lù àwọn?

Àwọn àkọlé mìíràn nígbà tí mò bá lo Distil-Whisper?

Bawo ni Distil-Whisper ṣe yatọ si awọn awoṣe STT miiran?

Mo le lo Distil-Whisper nípa API?

Ṣe mo le ṣe Distil-Whisper lori sévà mì?