Corpora Blog
Engineering guides on YouTube video and audio datasets, yt-dlp at scale, proxies and bandwidth planning for multimodal AI training data.
-
How to Build a YouTube Video Dataset for Multimodal Training
Build a YouTube video dataset for multimodal training: metadata-first discovery, filtering, yt-dlp downloads, scene splitting, dedup, ASR, shards and lineage.
-
Download YouTube Video Datasets: Panda-70M, HD-VILA, InternVid
How to download YouTube video datasets like Panda-70M, HD-VILA-100M and InternVid from IDs and timestamps: tooling, clip cutting, storage math and link rot.
-
yt-dlp Proxy Guide: --proxy, SOCKS5, Sessions and Workers
How to use a yt-dlp proxy: --proxy URL format, socks5 vs socks5h, env vars, config files, Python API, sticky sessions, aria2c and fixing 403/407 errors.
-
yt-dlp "Sign in to confirm you're not a bot": How to Fix It
Fix yt-dlp "Sign in to confirm you're not a bot" plus 403/429 errors: update, JS runtime, PO tokens, sleep flags, cookies and IP reputation at scale.
-
Bandwidth for Large-Scale Video Downloads: 100 TB to 5 PB/Month
How much bandwidth for large-scale video downloads? Formula and tables for 100 TB, 1 PB and 5 PB per month, plus utilization, storage and cost math.