Transcribe Audio & Video — Japanese, Chinese, Korean & 90+ languages

Turn your own audio and video files (meetings, interviews, podcasts, lectures, voice memos) into text, timestamped segments, SRT or WebVTT subtitles — with clean output for Japanese, Chinese and Korean.

• Properly punctuated Japanese and Chinese (、。 / ，。) — not one long unpunctuated line
• No stray spaces between CJK characters
• Timestamps, SRT and VTT subtitles
• 90+ languages with Whisper, plus optional translation to English
• Vocabulary hint for names and product terms
• Pay per audio minute — failed downloads and unreadable files are free

What can it be used for?

• Meeting minutes and interview transcripts (Japanese / Chinese / Korean / English)
• Subtitles for your videos (SRT / VTT)
• Turning podcasts and lectures into searchable text or LLM / RAG input
• Translating foreign-language audio into English text

Example output (Japanese test audio, 31 s)

本日は、新しい文字起こしサービスのテストを行います。音声を正しく文字に変換できるかどうかを確認するため、いくつかの文章を読み上げます。来週の月曜日から、事務所の受付時間が午前9時から午後5時半までに変わります。…

Input

Field	Description
audioUrls	Direct links to audio/video files (MP3, M4A, WAV, OGG, FLAC, MP4, MOV, WebM, ...). Only these URLs are downloaded.
language	auto or a language code (ja, zh, ko, en, ...). Setting it is more reliable for short files.
translateToEnglish	Output an English translation instead
prompt	Names or terms in the audio, to improve spelling
includeSegments / includeSrt / includeVtt	Output options
maxDurationMinutes, maxFileSizeMb, downloadTimeoutSecs	Limits

{
    "audioUrls": ["https://example.com/meeting.m4a"],
    "language": "ja",
    "includeSrt": true
}

Output

{
    "url": "https://example.com/meeting.m4a",
    "status": "success",
    "language": "ja",
    "model": "whisper-small",
    "durationSec": 31.3,
    "billedMinutes": 1,
    "charCount": 174,
    "text": "本日は、新しい文字起こしサービスのテストを行います。…",
    "segments": [ { "start": 0, "end": 5, "text": "本日は、新しい文字起こしサービスのテストを行います。" } ],
    "srt": "1\n00:00:00,000 --> 00:00:05,000\n本日は、新しい文字起こしサービスのテストを行います。\n\n2\n...",
    "realtimeFactor": 1.11
}

Pricing

Event	Price
Audio minute (per started minute of transcribed audio)	$0.05

Example: a 30-minute meeting costs $1.50. Failed downloads and unreadable files are not charged. You can cap the cost of a run with the maximum charge setting in Apify.

Speed

Transcription runs on CPU with Whisper (small model): roughly 1x realtime at the default 8 GB memory (a 10-minute file takes about 10 minutes). More memory = more CPU = faster.

Limitations

• Speaker labels (diarization) are not included.
• Very noisy audio or heavy background music reduces accuracy.
• Only your own files / files you have the right to process — the Actor does not download from YouTube or other platforms.

Privacy

Only the URLs you provide are downloaded. Audio is processed inside the run and deleted when the file is done; the transcript is written only to your own run's dataset.