Communitygithub.com

GML-MMGroup/ClipTalk

Generate local motion-graphics clips such as title cards, animated intros, visual explainers, and cover motion using the built-in HTML/Playwright/FFmpeg pipeline without external APIs.

What is ClipTalk?

ClipTalk is a Claude Code agent skill that generate local motion-graphics clips such as title cards, animated intros, visual explainers, and cover motion using the built-in HTML/Playwright/FFmpeg pipeline without external APIs.

Works with~Claude Code~Codex CLI~Cursor
npx skills add https://github.com/GML-MMGroup/ClipTalk/tree/HEAD/skills/cliptalk-local-motion-renderer

Ask in your favorite AI

Open a new chat with this agent skill pre-loaded.

Documentation

ClipTalk Local Motion Renderer

Use this skill when the user asks for Remotion-like, HyperFrames-like, HTML video, animated text, title-card, intro, dynamic cover, or visual explainer assets that can be rendered locally.

Hard boundaries:

  • Do not call cloud rendering, TTS, music, image-generation, or video-generation APIs.
  • Produce a separate local motion clip unless the user also asks to merge it into an existing cut; when they do, use compose_motion_intro to create a new output version.
  • Keep the output as an auditable preview artifact; formal delivery still needs the delivery-export skill.
  • Prefer the project’s ClipTalk visual identity, dark UI palette, readable Chinese typography, and safe-area margins.

Execution:

  1. Inspect the workspace only to confirm the current task and any requested aspect ratio.
  2. Render the motion clip with render_motion_graphics.
  3. If the user asked for a motion intro on the current cut and the workspace has an output, compose it with compose_motion_intro.
  4. Run run_delivery_qc on the generated local clip or composed output.

Related Skills