Se-arab e mahala
Ho ngola livideo le li-audio tsa Se-arab ho ea ho li-text ka AI. Habonolo, ka nepo, le mahala.
Ho sebetsa joang
- E_la ho Motsamaisi oa ho ngola oa Free.ai
- Kopitsa faele ea hau ea audio kapa video ea Se-arab
- AI ea rona e fumana Se-arab ka ho toba'me e e ngola
- Kopitsa transcript ea hau e le tlhaloso kapa li-subtitles tsa SRT
Se-arab Likarolo tsa ho ngola
- ✓E entsoe ke faster-whisper (license ea MIT)
- ✓Ho bona ka ho toba puo ea Se-arab
- ✓E tšehetsa MP3, WAV, MP4, M4A, FLAC, le tse ling
- ✓Timestamps le ho romelloa ha lihlooho tse ka tlase (SRT)
- ✓Ha ho na liphelelo tsa boholo ba faele ka li-plans tse lefelloang
- ✓Secha le se sireletsehileng -- lifaele li tlosoa ka mor'a ho sebetsana
Lintlha tsa puo
| Senyesemane | Se-arab |
| ISO Code | ar |
| Mofuta oa AI | faster-whisper |
| Theko | E le mahala |
Li-languages tse ling
Bona Li-Languages Tseo kaofelaLipotso tse ka etsahalang
Whisper large-v3-turbo e sebetsana le Se-arab ka ho feletseng - 7-15% palo ea liphoso tsa mantsoe ka molumo oa benchmark. E-ba le liphetoho tse sa tloaelehang ka li-entities tse bitsoang, lipalo, le mantsoe a tekheniki a thata; karolo e kholo ea transcript e tla ba ea nepahetseng.(Tier B, 7-15% word error rate ka lihlopha tsa benchmark - re phatlalatsa li-tiers tsa WER tse tšepahalang ho feta li-claims tsa thekiso.)
Ha ho joalo - Se-arab transcription e nka ho tloha letsatsi le letsatsi token pool mahala pele. Audio theko ka 50 tokens ka metsotsoana, ka hona pool letsatsi le letsatsi Anonymous e koahela lihora tse ling tsa audio ka letsatsi. Signed-in akhaonteng fumana 30,000-token letsatsi le letsatsi pool. Past hore, transcription ke pay-as-you-go, le token top-ups ho tloha $1.
Arabic e phetheloa ka boemo ba Modern Standard Arabic (MSA) ka ho tloaelo. Litlhaku tsa Egyptian, Levantine, Gulf, le Maghrebi li tšoauoa empa li ngotsoe ka ho ngola ka MSA - Whisper ha e romanese kapa e boloka ho ngola ka puo e itseng. Bakeng sa litaba tsa MSA tse hloekileng/lingoliloeng tsa audio, o tla hloka ho nahana ka ho nepahala ha tier-B; Maghrebi e boima kapa litlhaku tsa Egyptian li tla tlatsa ho tlase.
MP3, WAV, M4A, FLAC, OGG, OPUS, le WEBM li amoheloa ka ho toba. Bakeng sa video (MP4, MOV, MKV) re tlosa li-track tsa audio ka lehlakoreng la mosebeletsi pele re li romella ho Whisper - ha u hloka ho fetola ntho efe kapa efe ka boeona. Pipeline e ts'oanang ntle le puo ea mohloli, ho kenyeletsoa Se-arab.
Anonymous uploads cap at roughly 500 MB per file. Signed-in accounts go up to 2 GB. Duration is not a hard limit - long files are chunked automatically (30-second windows with overlap) and stitched back into a single transcript with continuous timestamps. Multi-hour Se-arab recordings (podcasts, full lectures, meetings) work fine.
Ho na le - mohlophisi diarization ke ka ho feletseng ka ho feletseng bakeng sa Se-arab transcripts tsohle. The output ke segmented joaloka Mohlophisi 1 / Mohlophisi 2 / Mohlophisi 3 le timestamps, kahoo lipotso, paneli lihlooho, le multi-party dibopeho tsa motheo ba tla khutla labelled. Diarization e tsamaea ka model e fapaneng le sebetsa ka tsela e tšoanang ka lipuo tsohle re tšehetsa.
E-na le hoo, re tla kenya URL ho /transcribe/youtube/ bakeng sa YouTube kapa /transcribe/podcast/ bakeng sa podcast feeds (Apple, Spotify, RSS). Re tla kenya audio, re e phethela ka Whisper le language=ar, ebe re khutlisa transcript le timestamps le li-labels tsa moqoqi. Lintlha tse tloaelehileng tsa Se-arab: livideo tsa litaba, lithapelo, lingoliloeng, le lipotso tsa bolumeli ka Se-arab ke li-workloads tse tloaelehileng ka ho fetisisa; kenya URL ea YouTube ho /transcribe/youtube/ kapa u romelle faele.
Whisper costs about 50 tokens per minute of audio, so a one-hour recording is ~3,000 tokens. Most users never spend anything - the free daily pool of 30,000 tokens covers short clips, voice notes, and one-off podcasts. Beyond it, transcription is pay-as-you-go, with token top-ups from $1.
Ee - li-timestamps tsa boemo ba segment (ka ~10-30 sekontiri ka 'ngoe) le boemo ba mantsoe li fumaneha. Lefatše la mantsoe ke la morao-rao bakeng sa ho romelloa kantle ha li-subtitles tsa VTT/SRT ka hona li-captions li synchronize line-by-line. Ka API beha timestamps="word" ka'mele oa kopo. Li-transcripts tsa Se-arab li khutlisoa ka ho ngola ha bona ho tloha ho le letšehali ho ea ho le letona'me li hlahisoa ka nepo ho mang kapa mang ea bonang RTL-aware (bahlolisani, Lentsoe, Google Docs).
Yeah. POST audio (multipart/form-data, field name "file") to /v1/transcribe/ with language=ar - or leave out the language parameter to let Whisper auto-detect. Returns JSON with the transcript, segments, timestamps, and speaker labels. Full reference and SDK snippets at /api/.
E-ea — ha transcription e felile, tobetsa ho fetolela kapa ho kopanya tekanyo ho /translate/. Se-arab e kopantsoe le lipuo tsohle tse ling tseo re li tšepang (200+). Bakeng sa lihora tsa kopano, transcript e fetisoa ka /summarize/; bakeng sa ho ngola ka letsoho, e romelle ho /voice/tts/ ho etsa molumo ka puo e loketseng.
Whisper e koetlisitsoe ka lihora tse likete tse likete tsa molumo oa lefatše la 'nete, ka hona e tobana le mollo oa morao-rao le ho ngolisoa ha boleng ba mohala ho Se-arab. Ho fihlela litlamorao tse ntle, fana ka molumo o hloekileng (mikrofono ea headset, ha ho na lebokose la molumo) - ka nako ena mollo o kopanya palo ea liphoso tsa baseline.Haeba transcript e sa khone ho sebelisoa, romella lengolo-tsoibila contact@free.ai le faele — re tla lefa tokens'me re sheba hore na e mong oa li-engine tse fapaneng o sebetsana joang le audio ea hau ka ho fetisisa.