YouTube Transcript Pro — Videos, Shorts, Playlists & Channels
Pricing
from $5.99 / 1,000 successful transcripts
Extract YouTube transcripts from videos, Shorts, playlists, channels, and search. Get text, timestamped segments, SRT, VTT, RAG chunks, and metadata. Supports language selection, auto captions, translation, and reliable proxy fallback. Pay only for successful transcripts.
Video URLs/IDs, playlist or channel URLs, @handles, or search:query entries. Defaults to a short public captioned video for a health check.
Optional additional list of video URLs or bare 11-character IDs.
[]Optional search terms that yt-dlp resolves into public video results.
[]Language priority list; exact match and locale-compatible variants are selected deterministically.
[ "en"]When enabled, use a manual track then an auto track only if none of the requested languages match.
Prefer human-created captions over auto-generated tracks when language priority allows.
If false, auto-generated tracks are not eligible for selection.
Optional target language code when the selected YouTube caption track supports translation.
Choose any of text, segments, SRT, VTT, or RAG chunks. All are outputs of one transcript event.
[ "text", "segments"]Also create timestamped RAG chunks, grouped near the selected chunk duration.
Target duration for RAG chunks; actual boundaries follow caption segments.
Optional lower bound for transcript segments; original timestamps are preserved.
Optional end timestamp; use with startSec or durationSec.
Optional duration used with startSec when endSec is not supplied.
Limit playlist/channel discovery; 0 means all discovered items.
Maximum results to resolve for each search query.
Include the channel Videos tab during channel discovery.
Include the channel Shorts tab during channel discovery.
Include the channel Streams tab during channel discovery.
Maximum wait per metadata or caption request.
Number of short retries for retryable per-video failures.
Retry NO_CAPTIONS only for videos published within 48 hours or recently-ended live replays with recent metadata.
Total upper bound for fresh-caption retries using finite 30/60/120-second backoff.
Direct uses no proxy; datacenter and residential use that group directly; auto tries direct, then datacenter, then optional residential.
Optional Apify Proxy settings for legitimate public-access reliability. Proxy mode controls the selected network group.
Allow residential proxy only after direct and datacenter network/bot failures in auto mode. Residential remains disabled by default.
Optional local cap on the number of confirmed successful-transcript events; 0 uses the Apify run charge limit only.
When prior state exists, skip items with a previously stored identical transcript hash.
Optional key suffix used for persistent resume and incremental state.
Skip previously confirmed successful transcripts under the selected state key.