Talking-head videos can have captions that appear right behind the speaker. This creates a professional and cinematic feel. The captions are not just plain text. They use a technique called occlusion where the speaker's body hides part of the text. This makes the text look like it is part of the scene.
There is a whole catalog of visual styles to choose from. You can pick a clean anchor rail or a dramatic neon sign effect. Every style is built from a set of rules so everything stays consistent. The pipeline handles transcription, matting, and compositing automatically.
Most explainer videos only need a few captions emphasized. The default Standard mode puts key words behind the speaker. For full cinematic captions use Cinematic mode. For wild VFX looks like stardust or ordnance use Theme mode.
Global
mkdir -p ~/.claude/skills/embedded-captionsProject
mkdir -p .claude/skills/embedded-captionsSource Repository
Video Editagentspace-so/runcomfy-agent-skills
Pick the perfect AI video editor for your style and needs
Video Inpaintingagentspace-so/runcomfy-agent-skills
Remove unwanted objects and clean up videos using AI inpainting across frames
Video Inpaintingruncomfy-com/skills
Remove objects from videos using simple prompts and smart AI models
Video Editruncomfy-com/skills
Smart video editing that picks the perfect AI model for your task
Video Inpaintingdoany-ai/skills
Remove objects, watermarks, and wires from videos with AI
Video Editdoany-ai/skills
Automatically choose the best AI model for your video edits on RunComfy
Hyperframesheygen-com/hyperframes
Build stunning videos using HTML animations captions voiceovers and transitions
General Videoheygen-com/hyperframes
Create custom video compositions for any length or format as a fallback