YouTube Transcript Corpus Audit & RAG Readiness
Pricing
from $6.00 / 1,000 transcript rag chunks
Extract public YouTube captions, audit transcript coverage, score RAG readiness, and create timestamped supporting chunks without double charging report mode.
YouTube watch, Shorts, embed, live, or youtu.be URLs.
[ "https://www.youtube.com/watch?v=dQw4w9WgXcQ"]YouTube channel URLs or bare @handles. The actor expands public channel pages into visible video IDs.
[]Optional public playlist URLs to expand into visible video IDs.
[]Preferred caption language code such as en, ja, or es.
Allow YouTube auto-generated captions when manual captions are unavailable.
Optional YouTube transcript translation target language code.
Maximum videos to process in one run.
Approximate maximum transcript characters per RAG chunk.
Approximate overlap characters carried from the previous chunk, segment-aligned.
Maximum transcript text characters to chunk per video.
Request timeout in milliseconds.
Public report tiers produce one decision-ready audit report. Use chunks only when you need legacy per-chunk output; watch summaries are planned and proof-gated.
Safety cap checked before report-tier charging. If the selected report price exceeds this cap, the actor returns a no-charge limit_reached row.
Report-mode safety limit. V1 emits one corpus report per run.
Return a no-charge sample report without external requests.
Remove duplicate video IDs after combining video, playlist, and channel sources.
Maximum chunk/warning rows to emit.
Choose whether to write rows to the dataset, post the final payload to webhookUrl, or both.
HTTPS endpoint used when delivery is webhook or dataset_and_webhook.