Podcast & Audio Transcription

Transcribes recordings and podcast feeds with speaker labels and timestamps, and saves SRT and VTT subtitles; only new episodes on a schedule.

Apify Store · coming soon

What it does

The Actor transcribes audio and video recordings, given as direct links or as podcast RSS feeds, into text with speaker labels and timestamps, and saves SRT and VTT subtitle files. In "only new episodes" mode it remembers which episodes of each feed it has done and transcribes only new ones.

Input

• Audio or video links: direct links to MP3, M4A, WAV, MP4, WEBM, OGG or FLAC files, up to 1 GB each. Links to web pages (for example a video site's page) don't work; use the file's own link.
• Podcast RSS feeds: the newest episodes with an audio or video file are transcribed.
• Episodes per feed: default 3, at most 100.
• Only new episodes: transcribe only episodes not done by an earlier run with the same state name. The memory is a key-value store named podcast-transcription-state in the user's own Apify account; a new state name starts over.
• Language: a two-letter code (en, de, es, fr, ...) or empty to detect it.
• Speaker labels: on by default.
• Maximum minutes per recording: default 240, at most 600.

Output

One dataset item per recording: link, feed, episode title, guid and publication date, status (ok, failed or skipped), duration, seconds transcribed, whether only the start was transcribed ("trimmed"), number of speakers, the text (one paragraph per speaker turn), the segments (start and end in seconds, text, speaker), links to the .txt, .srt and .vtt files in the run's key-value store, and an error message for failed or skipped items. The OUTPUT record has the run's totals. File links need the user's Apify token, like any record in their storage.

Pricing

Pay per event: one "audio minute" per started minute transcribed, plus Apify's small start event per run. Speaker labels, timestamps and subtitle files are included. Failed recordings are not charged.

Free Apify plan

On Apify's free plan a run transcribes up to 10 minutes in total; longer recordings are transcribed from the start up to that limit and marked "trimmed", and recordings beyond it are skipped with a note. Paid Apify plans have no such limit.

Troubleshooting

• "not an audio or video file that can be read": the link returned something else, often a web page; use the direct file link (for podcasts, the feed's enclosure link).
• "download failed: HTTP 403" or "404": the host refused the download or the file is gone.
• Speaker numbering restarts after an hour: recordings longer than about 75 minutes are transcribed in one-hour parts.
• Weak text: music, noise and people talking over each other lower the quality; setting the language helps for short clips.