MCP serverFor builders

Claude Video Vision

Ask Claude Code about a video's scenes and audio

jordanrendric/claude-video-vision1.4kstars on GitHubUpdated 14 days agoMIT licence

Where it can help

The plugin extracts frames and audio information so Claude can help inspect specific moments, summarize a clip or answer questions about what appears on screen.

Based on the project's documentation; not yet tested by Mostly AI.

Creators and developers using Claude Code who want to inspect a video

Where people talk about it

How to start

For builders. A longer technical setup: code, a server or Docker.

What's an MCP server?

MCP (Model Context Protocol) is a standard way to plug an AI assistant into other apps. The server is the plug: once it's added, the assistant can read from or work in that app when you ask.

1Check the project requirements and install the plugin using its official guide.
2Run its setup wizard and choose a local or cloud audio backend.
3Ask about a short video and check its answer against the original.

Open the project setup guide

Before you start

Claude Code
Node.js 24 or newer
FFmpeg
A local Whisper setup or a supported audio API

The catch: Cloud audio backends can send audio to a provider and may cost money. Local processing has its own setup and model download. Outputs can miss details.

For short videos. Checked Oct 28, 2026; stars and dates come from GitHub and refresh twice a day.

Claude Video Vision

Where it can help

Where people talk about it

More like this

Remotion Agent Skills

YouTube Full

HyperFrames