Sanskrit oo bilaash ah
Ku qor Sanskrit audio iyo video in qoraalka la AI. Fast, sax ah, iyo bilaash ah.
Sida ay u shaqeyso
- Tag Free.ai Qore
- Soo deji faylkaaga Sanskrit ee maqal ama fiidiyow
- AI-keena ayaa si otomaatig ah u ogaada Sanskrit oo u soo gudbiya
- Soo dejiso qoraalkaaga sida qoraal ama subtitles SRT
Sanskrit Faa'iidooyinka Soo-qaadista
- ✓Waxaa soo saaray faster-whisper (MIT licensed)
- ✓Ogaanshaha Sanskrit luqadda oo otomaatig ah
- ✓Taageer MP3, WAV, MP4, M4A, FLAC, iyo in ka badan
- ✓Soo dejinta taariikhaha iyo subtitles (SRT)
- ✓Ma jiro xaddidaadyo faahfaahsan oo ku saabsan qorshayaasha la bixiyo
- ✓Khasaaro iyo ammaan -- faylasha waa la tirtiri doonaa ka dib marka la xaliyo
Faahfaahinta Afka
| Afka | Sanskrit |
| Koodka ISO | sa |
| Nooc | Faster-whisper |
| Qiimaha | Bilaash |
Afaf kale
Afaf DhammaanSu'aalaha La Weydiiyo
Sanskrit waa af yar-maalgelinta u ah Whisper — weyn-v3-turbo fadhiya kor ku 25% qiimaha qalad eray, qaar ka mid ah si fiican kor. Qoraalka waa waxtar u leh raadinta iyo gist laakiin ma aha in la daweeyo sida daabacaadda diyaar u ah. Haddii ay ka sarreeya-dejinta mashiinka u noqon doontaa la heli karo Sanskrit waxaan ku xiran tahay oo si otomaatig ah.(Tier D, over 25% word error rate on benchmark sets - waxaan daabacnaa WER tier run ah ka badan dacwadaha suuqgeynta.)
Haa — Sanskrit qoraalka ka soo qaadaa maalin kasta free token pool hore. Audio ku kacaysaa ku saabsan 50 tokens daqiiqadii, sidaas darteed pool maalin kasta aan la aqoonsan oo ka kooban dhowr saacadood oo audio maalin kasta. Accounts Signed-in helaan pool weyn iyo 10,000 tokens diiwaangelinta. Hore, $ 1 iibsato 750,000 tokens (~ 250 saacadood oo audio).
Sanskrit transscripts waxaa lagu celiyaa UTF-8 caadiga ah oo leh afka caadiga ah ee afka.
MP3, WAV, M4A, FLAC, OGG, OPUS, iyo WEBM waxaa si toos ah loo aqbali doonaa. Video (MP4, MOV, MKV) waxaan soo saarno audio track server-xagga ka hor inta uusan u soo diri in ay ku qayliyo — ma u baahan tahay inaad waxba ku beddelaan adiga. Same pipeline kastoo afka asalka ah, oo ay ku jiraan Sanskrit.
Anonymous soo dejisan cap at ku dhowaad 500 MB file kasta. Accounts Signed-in u socdaan ilaa 2 GB. Muddada ma aha xad adag - files dheer ayaa si otomaatig ah chunked (30-second windows la isku dheelitirka) iyo dib u soo jiidasho in qoraalka kaliya oo leh timestamps sii socda. Multi-saacadaha Sanskrit diiwaangelinta (podcasts, akhris buuxa, kulan) shaqo fiican.
Haa — hadalka diarization waa on by default for every Sanskrit transcript. soo saarka waa qaybo sida hadalka 1 / hadalka 2 / hadalka 3 oo leh timestamps, sidaas daraaddeed wareysi, wada hadalka guddiga, iyo kulankii xisbiyada badan ee dib u yimaadaan la yiraahdo. Diarization ku socda qaabka kala duwan oo u shaqeeya sida isku mid ah oo dhan luqad kasta oo aan taageerno.
Yes — paste the URL into /transcribe/youtube/ for YouTube or /transcribe/podcast/ for podcast feeds (Apple, Spotify, RSS). We download the audio, run it through Whisper with language=sa, and return the transcript with timestamps and speaker labels. Typical Sanskrit content: akhris, wareysi, qoraalo cod, iyo YouTube content in Sanskrit dhammaan shaqada — paste URL in /transcribe/youtube/ ama soo dejisan faylka si toos ah.
Wisper ku kacaysaa ku saabsan 50 tokens daqiiqad kasta oo audio, sidaas darteed hal saac oo diiwaangelinta waa ~ 3,000 tokens. $ 1 iibsato 750,000 tokens, taas oo ka shaqeysa ilaa ku dhowaad 250 saacadood oo audio dollar kasta. isticmaalayaasha badankoodu ma bixiyaan waxba - biyaha maalinlaha ah ee bilaashka ah waxay ka kooban tahay fiidiyowyo gaaban, qoraallo cod, iyo podcasts hal-off.
Haa — labada qaybta-heerka (oo dhan ~ 10-30 ilbiriqsi) iyo erayo-heerka timestamps la heli karo. Erayo-heerka waa default for VTT / SRT subtitle dhoofinta si ay u subtitles sync line-by-line. On the API set timestamps ="word" in the request body. Sanskrit transscripts waxaa lagu celiyaa UTF-8 caadiga ah oo leh afka caadiga ah ee afka.
Haa. POST audio (multipart / form-data, magaca goobta "file") in / v1 / transcribe / la language = sa — ama ka tag parameter afka si ay u oggolaadaan Whisper auto-ilaalinta. JSON la soo celin, qaybaha, timestamps, iyo hadalka labels. Full soo jeedinta iyo SDK snippets at / api /.
Haa — mar haddii qoraalka la dhammaystiro, guji Translate ama ku dheji qoraalka /translate/. Sanskrit waa isku mid oo leh luqado kale oo dhan oo aan taageerno (200+). Kulanka daqiiqadaha ah ee qoraalka ka gudba /summarize/; duubista u dir /voice/tts/ si aad u hesho maqalka afka la doonayo.
Tababarka qaylada ee Whisper ka caawisaa yar ee heerkan - caqabadda waa in tirada Sanskrit audio Whisper arkay tababarka, ma aha qaylo. Clean studio audio weli ku garaacaan audio qaylo, laakiin midna ma gaadhi doono saxnaanta aad ka heli lahaa on a high-ressource afka.Haddii qoraalka dib u soo laabto aan la isticmaali karin, email contact@free.ai la faylka — waxaan ku soo celin doonaa calaamadaha iyo eegno haddii ay jiraan matoorro kala duwan oo aad audio ka fiican.