Start free

Transkrip fail audio panjang. Sehingga 10 jam.Tiada tamat masa.

Kebanyakan alat gagal secara senyap melebihi 2 jam atau 2 GB. Saluran paip kami membahagikan fail panjang ke dalam lorong pemprosesan selari, memasang semula dengan cap masa yang sejajar, dan menyampaikan satu transkrip yang koheren — label penutur utuh. Pro dan Business kedua-duanya menerima sehingga 10 jam setiap fail. Business menambah had saiz lebih tinggi (5 GB vs 2 GB) dan barisan keutamaan.

Drop a file, or pick one

MP3 · WAV · M4A · MP4 · MOV · MKV · OGG · OPUS · FLAC · WEBM — up to 100 MB anonymously

Paste a link, we’ll fetch the audio

YouTube · TikTok · Vimeo · Twitter · SoundCloud · Spotify · 50+ more

Record straight from your browser

Sign up takes 30 seconds — recording opens right after, in the dashboard.

No card required~90s per 60-min fileSRT · VTT · DOCX · TXTFiles auto-deleted in 24h

↓ Watch what happens

Audio in. Transcript out.

Drop a file, paste a URL, or record live — text appears back with speaker labels and timestamps. Same pipeline whether the source is a file, a URL, or your microphone.

Microphone · liveREC 00:07.41
en-US auto-detected16 kHz mono
~90s
Transcript · streaming2 speakers · 47:08
S1

Thanks for making the time. I want to start with framing — what was the original hypothesis behind the project?

S2

Honestly, maybe forty percent. The shape held — the mechanics underneath had to be rebuilt almost entirely.

S1

What flipped it for you? Was there one customer call, one piece of data

~95% accuracy on clean audioSRT · DOCX · TXT · JSON

↓ This is the dashboard

This is what loads when the job finishes.

Same layout as the real dashboard — Summary, full Transcript, Speakers tab, Exports. Key points and action items extracted automatically. Auto-tags on every job.

Try it on your own file — it's free

Three real options · honest comparison

Built-in dictation, AI speech-to-text, or a human transcriber.

Three legitimate ways to get text from spoken audio in 2026. Each is best for different work. Honest numbers below — no claim that AI matches a professional human transcriber on hard audio.

Option 01

Built-in dictation

Live, free, on-device. Handy for short bursts while you talk.

Accuracy · clear English~85%
Speaker separationNo
TimestampsNo
Languages~30
60-min fileLive only
CostFree
Best forShort voice notes while you walk. Hands-free messaging. Dictating an email at a stoplight.
Option 02

AI speech-to-text

~30× faster than realtime. 100+ languages. Speaker labels. The sweet spot for most work.

Accuracy · clear English95%+
Speaker separationYes (Pro+)
TimestampsPer sentence
Languages100+ auto
60-min file~90 s
Cost · per min$0.03
Best forInterview recordings · meeting notes · podcast transcripts · YouTube subtitles · batch jobs · API automation · field journalism.
Option 03

Human transcriber

Gold-standard accuracy when the audio is hard or the stakes are legal.

Accuracy · clear English98–99%
Speaker separationManual
TimestampsPer turn
LanguagesPer transcriber
60-min file4–8 hours
Cost · per min$1–3
Best forLegal depositions · medical dictation · archival oral history · anything where 95%+ accuracy is non-negotiable.

Built-in dictation figures from public iOS / Android speech API benchmarks. Human transcriber rates from US/UK industry surveys 2024–2025.

Accuracy · real-world numbers

95%+ on clear English. It holds up on real-world recordings too.

Modern transcription reaches 95%+ word accuracy on clear English at 128 kbps and above, comparable to a human transcriber on the same recording. The audio coming in sets the ceiling — cleaner source, cleaner transcript. The breakdown below covers the recordings we actually see in production.

97%+
Clean studio audio

USB or studio microphone in a treated room. Single speaker at conversational distance. The headline number.

95%+
Clear English at 128 kbps+

Podcast masters, interview recordings, well-mic'd meetings. The sweet spot for most professional work.

93%
Real-world podcast

Field-recorded interviews, podcast episodes at 64–128 kbps, multi-speaker recordings. Usable for editorial without a review pass.

91%
Meeting room recording

Ceiling mic, omnidirectional capture, mild reverb, multiple speakers at distance. Plan a rename pass on the speaker chips.

Common questions

7 things people ask about this.

01Apakah saiz dan panjang fail maksimum?+
Business: 5 GB setiap fail, sehingga 10 jam setiap rakaman. Pro: 2 GB setiap fail, sehingga 10 jam setiap rakaman. Percuma: 100 MB setiap fail, sehingga 30 minit setiap fail, 30 minit jumlah sebulan. Semua pelan menerima set format yang sama: MP3, WAV, M4A, FLAC, OGG, MP4, MOV, WEBM, MKV.
02Adakah rakaman saya akan tamat masa semasa muat naik atau pemprosesan?+
Tidak. Fail melebihi 2 GB muat naik terus ke storan melalui URL pra-tandatangan — permintaan tidak pernah tamat masa dalam tab pelayar. Pemprosesan menggunakan barisan kerja yang berterusan: walaupun anda menutup pelayar, kerja diteruskan dan anda menerima pemberitahuan e-mel apabila transkrip sedia.
03Adakah saya perlu membahagikan fail saya sebelum muat naik?+
Tidak. Saluran paip membahagikan fail secara dalaman kepada ketulan bertindih, memprosesnya secara selari, dan memasang semula satu keluaran dengan cap masa yang sejajar. Pemecahan manual hanya diperlukan jika fail anda melebihi had pelan.
04Adakah diarisasi penutur berfungsi merentas fail 10 jam?+
Ya. Diarisasi dijalankan pada keseluruhan fail selepas semua ketulan selesai, bukan secara bebas pada setiap ketulan. Ini bermakna SPEAKER_01 pada jam 1 adalah orang yang sama dengan SPEAKER_01 pada jam 9. Hanyutan biasanya di bawah satu salah atribusi sejam dalam audio terkawal.
05Seberapa cepat ia akan memproses fail saya?+
Kira-kira 4× lebih cepat daripada masa nyata — jadi rakaman 4 jam mengambil kira-kira 60 minit. Fail yang sangat besar (>6 jam) mungkin beratur di belakang kerja lain semasa waktu puncak; pemprosesan luar puncak adalah lebih cepat secara konsisten.
06Bolehkah saya mengumpulkan berbilang fail panjang melalui API?+
Ya. POST setiap fail sebagai kerja berasingan ke /api/v1/jobs dan nyatakan URL webhook. Kerja diproses secara selari (tertakluk kepada had serentak pelan). Business menjalankan sehingga 50 kerja serentak; Pro sehingga 20. Penghantaran melebihi had aktif diletakkan sebagai `pending` dan dinaikkan taraf secara automatik apabila slot dibebaskan — anda tidak akan dihadkan kadar pada penghantaran.
07Adakah cara percuma untuk menguji sebelum menaik taraf?+
Ya — peringkat Percuma memberikan anda 30 minit sebulan, tiada kad diperlukan. Gunakannya dengan petikan 30 minit yang mewakili rakaman panjang anda untuk memeriksa ketepatan sebelum komited pada pelan berbayar.

Drop something in. See what comes out.

Muat naik sehingga 5 GB, 10 jam panjang. Tiada pemecahan, tiada tamat masa, label penutur sepanjang.

Mulakan transkripsi percuma