Get transcripts from YouTube videos, whole channels or playlists in one run. JSON with timestamps, SRT, VTT or plain text. 2 USD per 1,000 transcripts, and videos without captions are free.
Error items now use one shared set of codes across all Vakitek YouTube actors. Each error item has the shape {error, input, message, retryable}. The full list is in the README, section Errors.
NO_TRANSCRIPT is now NO_CAPTIONS when the video has no captions, or BLOCKED when YouTube blocked us (a block is never reported as "no captions" anymore)
VIDEO_UNAVAILABLE is now NOT_FOUND (video removed) or BLOCKED (video unplayable for us)
CHANNEL_NOT_FOUND is now NOT_FOUND
PLAYLIST_NOT_FOUND is now NOT_FOUND
LOGIN_REQUIRED is now BLOCKED, or AGE_RESTRICTED / PRIVATE / MEMBERS_ONLY when YouTube says why
The previous code is kept in a legacyError field, so a pipeline that matches on the old value keeps working while you migrate.
retryable: true tells you a rerun may succeed (block, rate limit, temporary upstream error).
Added
timeBudgetSecs input (default 240, 0 = no limit): the run returns what is ready before the limit, so n8n, Make and run-sync calls no longer time out at 300 seconds. Inputs not finished in time come back as a free TIME_BUDGET_EXCEEDED item.
Reliability section in the README: public success rate over the last 30 days, refreshed weekly.
"Use it from n8n, Make or an API" section in the README.