Conveyor belt for hauling YouTube videos to bilibili and AcFun

What it does

Y2A-Auto is a Python-based pipeline that ingests YouTube videos and republishes them on bilibili and AcFun with minimal human intervention. It handles downloading, optional speech-to-text via Whisper or Voxtral, subtitle translation and quality checks, AI-generated titles and tags, content moderation through Alibaba Cloud, and final upload. A built-in web dashboard lets you review tasks, manage settings, and monitor YouTube channels for new fodder.

The interesting bit

The project treats cookie management and subtitle hygiene as first-class problems rather than afterthoughts: it syncs YouTube cookies via CookieCloud, filters filler words and hallucinated text from subtitles, and can brute-force-lock the admin panel. That obsession with the boring glue—session persistence, punctuation normalization, path-traversal protection—is what keeps an unattended reposting bot from falling over.

Key highlights

• End-to-end automation from yt-dlp download to platform upload, with optional human approval gates.
• Subtitle transformation engine that splits long lines, normalizes punctuation, filters filler words, and detects hallucinated text before burning captions into the video.
• Hardware-accelerated transcoding with HEVC/H.265 preference and automatic fallback to H.264.
• YouTube channel and keyword monitoring that automatically enqueues new videos.
• Security features including password protection, brute-force lockout, session timeouts, and path-traversal guards.

Caveats

• The documentation and web interface are in Chinese, and integrations are tightly bound to Chinese platforms (bilibili, AcFun) and services like Alibaba Cloud and WeCom.
• Much of the pipeline is orchestration around yt-dlp, FFmpeg, and Whisper/Voxtral, so stability tracks those upstream dependencies and platform cookie formats.

Verdict

Worth a look if you operate channels that curate or mirror YouTube content onto bilibili or AcFun and want to automate the drudgery. Skip it if you are not targeting those platforms or if you expect a generic video workflow tool.