← all repositories
Norsico/Video-Materials-AutoGEN-Workstation

A one-man TikTok factory, held together with YAML and honesty

A Chinese-language web UI that wires Gemini, TTS, and image gen into a project manager for short-form video assembly lines.

Video-Materials-AutoGEN-Workstation
Velocity · 7d
+7.6
★ / day
Trend
steady
star history

What it does This is a browser-based workstation for churning out short videos. You create projects from templates, then lean on Gemini for scriptwriting, a TTS engine for emotional voiceover, and image models for visuals. Everything—scripts, audio, subtitles, stills—gets tracked per project through a card-based UI. The author is admirably blunt: it is “mainly for management,” not a magic viral-video button.

The interesting bit The README contains a rare warning: the author lost their own n8n workflow files, so the copywriting feature is effectively a prompt template you wire up yourself. Docker deployment is documented but flagged “currently has bugs.” The whole thing runs on Node, calls local/remote AI services via YAML config, and the containerized version cannot pop a file manager—paths are returned as text strings instead. It is held together with env.yaml and realistic expectations.

Key highlights

  • Project-centric UI: cards track output directories, creation time, and deletion for each batch run.
  • Split-track editing: swap images, subtitles, or audio independently and preview before export.
  • TTS batching: feed multiple scripts with emotion tags and get voiced tracks back.
  • ASR subtitle extraction: reverse-engineered interface for pulling transcript files, code credited to another author.
  • Image pipeline: prompt management for characters/scenes, reference uploads, aspect ratio locks, and history for reuse.

Caveats

  • Docker setup is explicitly marked buggy; source deploy (Node + npm) may be more reliable.
  • Several features require external wiring: n8n for copywriting workflows, another repo for subtitle fetching, and a local NanoBanana/AIStudio proxy for image generation.
  • The author notes the tool is less “practical” than it looks; content strategy and virality are still on you.

Verdict Worth a spin if you are already running a Chinese short-video operation and want a structured wrapper around Gemini + TTS + image gen. Skip it if you need a polished, end-to-end turnkey pipeline or do not read Chinese.

heatdrop uses Google Analytics to see which pages get read — nothing else. Your call. How we handle data.