Distil-Whisper large-v3
Free.ai (self-hosted)
·
stt
·
~500 tokens per minute
Distil-Whisper large-v3 iku a model swara-ka-teks dibangun déning HuggingFace. Strongest at Real-time transcription, large-volume batch STT.. Self-hosted on Free.ai GPUs — runs free against your daily token pool (500 tokens saben menit). Released under MIT — commercial use permitted on Free.ai.
Nganggo liwat API
REST API kang kompatibel karo OpenAI. Janji kunci lan waca model iki ing sekon.
curl -X POST https://api.free.ai/v1/stt/ \
-H "Authorization: Bearer sk-free-..." \
-H "Content-Type: application/json" \
-d '{"model":"distil-whisper-large-v3","audio_url":"https://..."}'Takon kang asring diajukake
Distil-Whisper large-v3 ngrekam swara kang diucakaké dadi teks. Ngunduh MP3, WAV, M4A, utawa file video lan Distil-Whisper large-v3 bakal ngembali transkripsi lengkap plus subtitle SRT/VTT pilihan karo timestamp.
Distil-Whisper large-v3 handles dozens of languages - Whisper-family models cover 90+, others vary. Pick "auto-detect" or specify the language for highest accuracy.
Kasalahan tembung rata-rata 5-10% ing audio Inggris kang bersih, 10-20% ing audio sing noise utawa akcented. Varian gedhe saka arsitektur sing padha luwih apik ing kasus hard - pilih luwih gedhé nalika audio ora rata.
Ya - saben segmen nglumpukaké stempel wektu wiwitan/akhiran. Eksport minangka SRT utawa VTT lan peta wektu langsung ing videomu.
Distil-Whisper large-v3 runs on our own GPUs against your daily free pool first; $5 → 200,000 paid tokens after that. About ~500 tokens per minute.
MP3, WAV, M4A, FLAC, OGG, plus video (MP4, MOV, WebM) - kita ngekstrak audio. Max 500 MB saben upload. File luwih dawa? Dipisahake karo /audio/cut/ utawa nggunakake /v1/stt/batch/.
Speaker diarization is a separate pass - toggle "diarize" on /transcribe/. Distil-Whisper large-v3 handles the transcription; diarization labels each segment with Speaker 1 / Speaker 2 / etc.
Ya - /batch/ nampi folder kanggé file audio. Saben transcript dipun-simpen ing /account/?tab=history kaliyan nama file asli. Kanggo perlindungan folder-tree, gunaka API.
Yes - POST your audio to /v1/stt/transcribe/ with model="Distil-Whisper large-v3". Returns JSON with text + segments + word-level timestamps. /api/ has the full reference.
Model self-hosted nyimpen audio ing GPU kita; premium ngliwati karo DPA. Audio dipatèni sawisé jendela share (24 jam anon, 7 dina mlebu). Kita ora latihan ing input sampeyan.
Ya - Free.ai nyedhiyani panggunaan komersial transkripsi. Sampeyan butuh hak kanggo audio sing sampeyan unggah (rekaman dhewe, materi lisensi, utawa isi kanthi persetujuan).
Faktor wektu nyata kira-kira 0.05-0.2× - podcast 60 menit ditranskripsi ing 3-12 menit. Model Premium asring rampung luwih cepet. Gunake tombol baris tunggu kanggo nutup tab.