claude-video
GitHub ↗
Video comprehension for agents
Let the
model
watch.
Frames and transcript, not guesswork