Fal Speech-to-Text
Free.ai
·
stt
·
~500 Ii-token minute
Fal Speech-to-Text yi an imodeli yokuthetha-ukubhaliweyo. Ihamba ngeendlela ezingaphandle — ~500 ii-token % 1 imizuzwana (50% yokuphawula ngaphezulu kwexabiso eliphezulu).
Sebenzisa nge-API
OpenAI-compatible REST API. Generate a key and call this model in seconds.
curl -X POST https://api.free.ai/v1/stt/ \
-H "Authorization: Bearer sk-free-..." \
-H "Content-Type: application/json" \
-d '{"model":"premium/speech-to-text","audio_url":"https://..."}'Imibuzo ebuzwa rhoqo
Fal Speech-to-Text iguqula umbhalo othethayo ube ngumbhalo. Layisha phezulu i-MP3, WAV, M4A, okanye ifayili yevidiyo kwaye Fal Speech-to-Text ibuyisela uguqulelo olupheleleyo kunye ne-SRT/VTT ekhethiweyo enezihloko zexesha.
Fal Speech-to-Text iphatha iileta eziliqela zeelwimi — Whisper-imodeli yosapho iquka i-90+, Parakeet iquka i-~25, ezinye zitshintsha. Khetha "ukukhangela ngokuzenzekelayo" okanye khankanya ulwimi ukuze uqinisekise umgangatho ophezulu.
Ixabiso legama-impazamo yi 5-10% kwisandi esicocekileyo sase-English, 10-20% kwisandi esinomsindo okanye esinomsindo. Iinketho ezinkulu zesi siqalo zisebenza kakuhle kakhulu kwiimeko ezinzima - khetha uninzi xa isandi sinzima.
Ewe - icandelo ngalinye liquka isiqendu sokuqala/sokuphela sexesha. Rhweba ngaphandle njenge-SRT okanye i-VTT kwaye ixesha lemaphu liyi-directly kwividiyo yakho.
Fal Speech-to-Text yinjini yokuguqulela ephezulu. Imalunga ne ~500-1,500 ye-token ngomzuzu wesandi. $1 = 750,000 ye-token.
MP3, WAV, M4A, FLAC, OGG, kunye nevidiyo (MP4, MOV, WebM) — sikhupha isandi. Ubukhulu be-500 MB nganye. Iifayili ezide? Yahlula nge /audio/cut/ okanye sebenzisa /v1/stt/batch/.
Ukwenza i-diarize yomthumeli kugqithiso oluhlukileyo - tshintsha "ukwenza i-diarize" kwi / transcribe /. Fal Speech-to-Text iphatha ukufakelwa; ukwenza i-diarize iphawula icandelo ngalinye ngeMthumeli 1 / Umthumeli 2 / njl.
Ewe — /batch/ ivuma isiqulathi seefayili zesandi. Ushicilelo ngalunye luya /account/?tab=history ngegama lefayili elisemthethweni. Ukugcina umthi wesiqulathi seefayili sebenzisa i-API.
Ewe — UTHENGA umculo wakho kwi /v1/stt/transcribe/ ngemodeli "Fal Speech-to-Text". Ibuyisela i-JSON ngombhalo + iinxalenye + i-timestamp yegama-level. /api/ inesiqendu esipheleleyo.
Iimodeli ezimkelweyo zigcina isandi kwi-GPU yethu; ipremiyamu idlula nge-DPA. Isandi sicinywa emva kwefestile yokuzonwabisa (24h anon, 7d ubhaliso-ngaphakathi). Asiqeqeshi kwingeniso yakho.
Ewe — Free.ai inikezela ngenkonzo yorhwebo yokusetyenziswa kweencwadi ezibhalwe ngesandla. Ufuna ilungelo lesandi olayishe (u Recording yakho, i-layout material, okanye imixholo enegunya).
Ixesha-lokwenyani limalunga ne-0.05-0.2× — ipodcast yemizuzu engama-60 iguqula kancinane kwimizuzu engama-3-12. Iimodeli zepremium zihlala zigqiba ngokukhawuleza. Sebenzisa iqhosha lofolo ukuvala i-tab.