youtube-transcriptlisted

# Youtube Transcript IRON LAW: NEVER give up after one method fails. Try the methods in order until one succeeds. ## Read this first — the 2025+ reality YouTube now aggressively gates the player/transcript endpoints. If you see **"Sign in to confirm you're not a bot"** (yt-dlp) or **`RequestBlocked`/`IpBlocked`** (youtube-transcript-api), the request was refused on **identity/IP grounds**, not format. On a datacenter IP, a VPN exit, or a rate-limited residential IP, the anonymous methods (1–3) will ALL fail — including a real browser and a valid PO token. Proven: a genuine Playwright browser gets a caption-stripped player response, and a correctly-minted bgutil PO token is still refused. **The only thing that reliably passes the wall is a logged-in session → go to Method C.** ``` Progress: - [ ] Step 1: Check dependencies - [ ] Step 2: Method 1 — youtube-transcript-api (fast; works on clean IPs) - [ ] Step 3: Method 2 — yt-dlp player clients (if M1 blocked) - [ ] Step 4: Method 3 — direct timedtext scrape (if M1–2 blocked) - [ ] Step 5: Method C — yt-dlp + cookies.txt ← RELIABLE when "not a bot" wall appears - [ ] Step 6: Save clean text and report which method worked ``` ## Step 1: Check dependencies ```bash python3 scripts/check-deps.py # installs: pip install --user youtube-transcript-api yt-dlp ``` ## Method 1 — youtube-transcript-api (try first) ```bash python3 scripts/yt-transcript.py <video_id_or_url> --lang en,ru --output out.txt ``` Fast when th