Claude video tool uses Gemini 2.5 Flash key, yt-dlp, and ffmpeg to watch videos
EXM7777 · x · 2026-07-26
A Claude video-watching hack routes prompts through a Gemini 2.5 Flash API key and uses yt-dlp plus ffmpeg to ingest YouTube, Loom, TikTok, and local files.
The tool, bradautomates/claude-video, first grabs free captions to avoid cost, then extracts scene-aware frames for tasks like competitor launch analysis, screen-recording debugging, fast talk summaries with timestamps, and stripping hype from product videos.
Related event: claude-video Plugin Enables Claude to Watch Videos(3 posts)→
More from coding & agent
- Live workshop promises a real-business GenAI + data science workflow with Python and agents — mdancho84 · 2026-07-26
- Codex computer-use handles Google Cloud Console setup tasks with less manual work — lucasmeijer · 2026-07-26
- One-Click Desktop AI Client: Run Claude & Codex Locally with Zero Inference Fees — jasonkneen · 2026-07-26
- Agent workspace lets humans and multiple sandboxes share one endpoint, one context and one compute pool — jasonkneen · 2026-07-26
- Built 47 Agents With n8n and Claude, Now Sharing 3 Prompts He Says Actually Work — Aiden_Tech_Ai · 2026-07-26
- A Sunday morning workflow: coffee, Codex, voice Jarvis, then build the cancer app — DeryaTR_ · 2026-07-26