YouTube Transcript API for Agents (pay per success)
Pricing
from $5.00 / 1,000 youtube transcript api — extract video transcripts to jsons
YouTube Transcript API for Agents (pay per success)
YouTube transcripts for AI agents: JSON/text/SRT with timestamps. Charged ONLY on delivered transcript - no captions or dead video = no charge. Residential proxy built in; reliable at 1,000+ videos/run.
Pricing
from $5.00 / 1,000 youtube transcript api — extract video transcripts to jsons
Rating
0.0
(0)
Developer
Omar Kurbanov
Maintained by CommunityActor stats
0
Bookmarked
2
Total users
1
Monthly active users
19 hours ago
Last modified
Share
YouTube Transcript API for Agents — pay per successful transcript
Get YouTube video transcripts with timestamps in JSON, plain text, or SRT — priced for AI agents and automation fleets: you are charged only when a transcript is actually delivered. No captions, private/deleted video, or any failure on our side = an error record and no charge.
Why this one
- Pay per success, not per attempt. Most transcript tools charge for the run
regardless of result. Here the
transcriptevent fires only on a delivered transcript. - Built for volume. Residential proxy routing is on by default — no "Sign in to confirm you're not a bot" failures that break datacenter scrapers. Reliable at 1,000+ videos per run.
- Always-fresh extractor. The YouTube backend churns constantly; this actor auto-updates its extraction engine on every run, so it doesn't silently rot.
- Timestamps included. Segment-level
start/durin JSON mode, proper SRT subtitle output, or clean plain text for LLM ingestion.
Input
{"videos": ["https://www.youtube.com/watch?v=jNQXAC9IVRw", "dQw4w9WgXcQ"],"format": "json","lang": "en"}
videos— YouTube URLs or 11-character video IDs, any mixformat—json(segments + timestamps),text(plain transcript),srt(subtitles)lang— caption language; manual captions preferred, auto-generated fallback
Output (dataset, one item per video)
{"video": "dQw4w9WgXcQ", "ok": true, "video_id": "dQw4w9WgXcQ","language": "en", "source": "manual","segments": [{"start": 18.8, "dur": 3.6, "text": "We're no strangers to love"}],"transcript": "We're no strangers to love ..."}
Failed videos come back with "ok": false and an error string — uncharged.
Use cases
RAG pipelines and LLM knowledge bases · content research and summarization at scale · podcast/interview mining · compliance and brand monitoring · dataset building · any agent workflow that needs "what was said in this video" as cheap structured data.
Also available over x402
The same engine runs as a pay-per-call HTTP API for x402-native agents
(USDC on Base, no account needed): https://transcriptd.fly.dev — same
pay-only-on-delivery guarantee.