Audio to Text Transcriber with Subtitles

Turn audio files into text: paste file links or upload files and get a clean transcript, timestamped segments and ready-to-use SRT and VTT subtitles. It reads MP3, WAV, FLAC and OGG files in 40+ languages, and can translate any of them into English.

Built for: podcast and interview transcripts, meeting and lecture notes, voice notes, subtitles for your own recordings, and feeding spoken content into AI, RAG and search pipelines from Make, Zapier or n8n.

What you get

• Clean transcript with paragraphs at natural pauses, plus word count and detected language.
• Timestamped segments: start and end time (in seconds) for every sentence or phrase.
• Subtitles: SRT and WebVTT text you can save as a file and load in a video editor or player.
• Word timestamps (optional): start, end and confidence of every word, for karaoke-style captions and search.
• Translate to English from any supported language in the same run.
• Two accuracy levels: Standard (fast, great for clear speech) and High accuracy (better with accents, noise and unusual words).
• Private by design: speech recognition runs inside your own Apify run. Your audio is never sent to an outside AI service.
• Fair billing: you pay per minute of audio. Files with no speech, unsupported files and failed downloads are free, and the run stops cleanly at your maximum cost per run.

How to use

1. Add audio file links (one per line) or upload files in Audio files.
2. Choose Accuracy and, if you know it, the Spoken language (Auto-detect works for most files).
3. Click Start, then download results as JSON, CSV, XLSX or HTML, or fetch them with the API.

curl -X POST "https://api.apify.com/v2/acts/ridgeline_apps~audio-transcriber/run-sync-get-dataset-items?token=YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"files": ["https://example.com/interview.mp3"], "language": "en"}'

Input

Field	Description	Default
files	Audio file URLs or uploaded files (MP3, WAV, FLAC, OGG)	required
accuracy	standard or high	standard
language	Spoken language code (en, es, de, ...) or auto	auto
task	transcribe (original language) or translate (to English)	transcribe
includeSubtitles	Add srt and vtt subtitle text	true
includeSegments	Add timestamped segments	true
includeWordTimestamps	Add words to each segment	false
maxFileSizeMb	Larger files are skipped (free)	200
maxMinutesPerFile	Longer files are skipped (free)	180

Output

{
  "url": "https://example.com/interview.mp3",
  "fileName": "interview.mp3",
  "status": "ok",
  "format": "mp3",
  "durationSeconds": 754.2,
  "billedMinutes": 13,
  "language": "en",
  "languageProbability": 0.98,
  "text": "Welcome back to the show. Today we're talking about...\n\nThanks for having me...",
  "wordCount": 1893,
  "segments": [
    { "start": 0.0, "end": 3.12, "text": "Welcome back to the show." }
  ],
  "srt": "1\n00:00:00,000 --> 00:00:03,120\nWelcome back to the show.\n...",
  "vtt": "WEBVTT\n\n00:00:00.000 --> 00:00:03.120\nWelcome back to the show.\n..."
}

Files that can't be processed come back with "status": "error" and a plain-English error. They are never charged.

Pricing

Pay per minute of audio, with no subscription. Each file is billed in whole minutes, rounded up (a 2:10 file counts as 3 minutes). High accuracy costs more per minute than Standard. Files with no speech and failed files are free. The exact prices are on the Pricing tab, and you can set a maximum cost per run in the run options.

Speed and run settings

On the default 4 GB memory, Standard transcribes about 8 minutes of audio per minute of run time; High accuracy about 3. For long files, raise the run Timeout in the run options (for example, a 3-hour recording on High accuracy needs about an hour). More memory gives more CPU and finishes faster.

FAQ

Which formats are supported? MP3, WAV, FLAC and OGG (Vorbis). M4A, AAC, Opus, WMA and video files are not supported: convert them to MP3 first. Unsupported files are reported as errors and not charged.

How accurate is it? On clear speech, transcripts are usually very accurate. Strong accents, background noise, crosstalk and names lower accuracy; use High accuracy and set the spoken language for the best results.

Does it label speakers? No. Speaker labels (diarization) are not included.

Can it transcribe a video-site or podcast page link? No. It needs a direct link to an audio file (or an upload). It does not download from video or streaming platforms.

Where does my audio go? Files are downloaded into your own Apify run and transcribed there. Nothing is sent to any third-party AI or transcription service.

Legal

Only transcribe audio you have the right to use, and follow the recording-consent laws that apply to you. You're responsible for the files you submit and how you use the transcripts. Not intended for medical records (PHI) or other regulated data that requires a signed processing agreement.

Built on open-source software: Whisper speech models by OpenAI (MIT), faster-whisper and CTranslate2 (MIT), miniaudio (MIT; public-domain decoders), ONNX Runtime (MIT), Hugging Face Tokenizers (Apache-2.0) and httpx (BSD-3-Clause).

Questions or a feature request? Open an issue on the Issues tab.