vea
por shawnshenopeninterx
Coding Agents & IDEsResumo
Video Editing Agent (VEA) for automated video processing, highlight generation, and editing.
Descrição
O que esta skill faz
VEA is a locally-run video editing service that indexes video content, selects clips using AI comprehension, and assembles edited videos with narration, subtitles, and background music. It runs as a FastAPI server on localhost and delegates AI tasks to Memories.ai, Google Gemini, and ElevenLabs.
Handles clip selection, narration, subtitles, and music mixing in a single local pipeline rather than requiring separate tools for each editing step.
Quando usá-la
- Creating highlight reels from long event recordings
- Adding AI voiceover to raw footage automatically
- Burning subtitles into clips for TikTok or Reels
- Reformatting landscape video to vertical aspect ratio
- Indexing a video library for AI-powered clip retrieval
Exemplo de fluxo de trabalho
User asks: Create a 1-minute highlight reel from my conference talk video
- 1
Start the VEA server locally and confirm it is running at localhost:8000
- 2
Index the source video via POST /video-edit/v1/index to enable AI comprehension
- 3
Present configuration options (narration, music, subtitles, aspect ratio) and wait for user confirmation
- 4
Send the highlight reel request to /video-edit/v1/flexible_respond with confirmed settings
- 5
Return the output video path from data/outputs/
A 1-minute MP4 highlight reel with AI-selected clips, voiceover narration, background music, and burned-in subtitles saved to ~/vea/data/outputs/
Requisitos
- MEMORIES_API_KEY (get at https://memories.ai/app/service/key)
- GOOGLE_API_KEY (Google Cloud Console)
- ELEVENLABS_API_KEY (ElevenLabs account)
- SOUNDSTRIPE_KEY (optional, Soundstripe account)
- FFmpeg installed on the system
Instalação
clawhub install shawnshenopeninterx/veaExecute o comando da CLI ou cole o prompt no Claude Code, Cursor, Codex ou no agente de código que você usa.