Podcast Transcription - RSS Episodes to Text
Pricing
from $4.80 / 1,000 audio minutes
Turn public podcast RSS episodes into text, TXT and optional SRT. Use publisher transcripts when available or transcribe audio by the minute. One newest episode per feed by default; failed inputs are free.
contentPack
Optional
Optional chapters, showNotes, socialPosts and translatedSubtitles.
contentPackError
Optional
Why the optional content pack was not made.
contentPackCharged
Optional
True when one content-pack event was charged.
input
Optional
The input line this row was made from (URL, uploaded file name, feed or dataset entry); on the one row of an input mistake it is the input field name (e.g. "language").
inputIndex
Optional
Position of the file in the input (0, 1, 2 ...), so rows saved out of order by parallel files can be mapped back to the input. Absent on rows not tied to one input item (input-error rows).
url
Optional
The audio or video URL processed. Absent on input-error rows.
mediaUrl
Optional
URL of the media file found on the page, when the input was a page (podcast episode, cloud share or social video) rather than a direct file link.
episodeTitle
Optional
Podcast episode title (podcast feed inputs); null when the feed gives no title.
videoTitle
Optional
YouTube video title (YouTube video inputs).
mediaTitle
Optional
Title of the media found on the page, when the page names it.
sourceKind
Optional
Kind of source the link is: youtube, gdrive, dropbox-shared, onedrive, tiktok, instagram, soundcloud, direct, and similar.
transcriptSource
Optional
"asr" (transcribed by speech-to-text) or "publisher" (the podcast's own transcript; no audio minutes).
success
Required
true when the file was transcribed: text holds the transcript and the requested files were saved. false on every free row: failures (errorType, error), files skipped at a limit (skipped:true) and valid files with no speech (noSpeech:true). Rows with success:false are never charged.
charged
Required
true when this row was billed: audio minutes (billedMinutes > 0), an AI summary (summaryCharged:true) and/or translated subtitles. false on every failure, skipped and no-speech row; a success:false row is always charged:false (invalid_input included).
skipped
Optional
true on files skipped before processing (errorType too_large or limit_reached: longer than maxMinutesPerFile, or over the minutes/spending limit). Free.
noSpeech
Optional
true when the file was valid but had no speech (silence or music only): a free no_text row with an empty text, counted as a delivered answer in SUMMARY.
errorType
Optional
Machine-readable reason on success:false rows: invalid_input (the input line could not be used: a free failure), no_text, too_large, limit_reached, blocked, timeout, network, rate_limited, http_error, download_failed, unsupported, unsupported_platform, private_media, not_found, no_media, multiple_media, private_link (a share link that is not public), needs_login (the link opened a sign-in page), quota_exceeded (the host's download quota: retry later), not_audio (a file that is not audio or video), folder_link, link_expired, link_invalid, failed. Absent on success rows.
error
Optional
Human-readable failure text on success:false rows, always ending with "(not charged)". Absent on success rows.
retryable
Optional
Present on every failure row: true when running the same input again may work later (busy host or service, network, blocked, rate limits), false for final causes (invalid_input, no_text, too_large, private_media, ...). Absent on success rows; never read an absent retryable as false.
language
Optional
Language of the transcript (ISO code such as "en"), detected or requested.
durationSeconds
Optional
Audio duration in seconds, when known.
billedMinutes
Optional
Audio minutes billed for this row; 0 on every free row. Absent on input-error rows.
wordCount
Optional
Words in the transcript; 0 on no-speech rows.
speakers
Optional
Number of distinct speakers, when speaker labels were requested.
text
Optional
The full transcript text of a successful row. Empty string on a no-speech row (noSpeech:true); absent on other failure rows.
segments
Optional
Timed transcript segments [{ start, end, text, speaker? }] with seconds from the beginning.
paragraphs
Optional
Transcript grouped into paragraphs [{ start, end, speaker?, text }].
summary
Optional
AI summary of the transcript (summarize on); billed as the summary event when present (summaryCharged:true).
summaryCharged
Optional
true when the AI summary was billed for this row.
srtUrl
Optional
Public URL of the row's SRT subtitles file in the run's key-value store (when subtitles were saved and the transcript is timed).
vttUrl
Optional
Public URL of the row's WebVTT subtitles file (same conditions as srtUrl).
txtUrl
Optional
Public URL of the row's plain-text transcript file (saveTranscriptFiles on and the transcript is not empty).
mdUrl
Optional
Public URL of the row's Markdown transcript file (same conditions as txtUrl).
meetingNotesUrl
Optional
Public URL of the Markdown meeting-notes file (meetingNotes on, charged audio).
translatedSubtitles
Optional
Translated subtitles (subtitleLanguages): one entry per language, each with its own success, language, languageName and srtUrl/vttUrl/cues/billedMinutes, or errorType/error when that language failed or was skipped (free).
keywordsFound
Optional
searchTerms terms found in the transcript.
keywordCounts
Optional
Mentions per search term, e.g. { "pricing": 3 } (terms with 0 mentions included).
keywordMatches
Optional
Every mention of a search term, in time order: [{ term, matched, start, end, time, link, text, speaker? }] (times null for untimed publisher transcripts).
transcribedAt
Optional
When this row was saved (ISO 8601 timestamp).
