Understand This Video For Me AI Plugin
Enables the server to accept a Loom video URL and extract timestamped key frames, a transcript using site captions when available, a contact-sheet image, and a combined timeline of visuals and speech for AI analysis.
Enables the server to accept a Vimeo video URL and extract timestamped key frames, a transcript using site captions when available, a contact-sheet image, and a combined timeline of visuals and speech for AI analysis.
Enables the server to accept a YouTube video URL and extract timestamped key frames, a transcript using site captions when available, a contact-sheet image, and a combined timeline of visuals and speech for AI analysis.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Understand This Video For Me AI PluginSummarize this YouTube video with timestamps: https://www.youtube.com/watch?v=dQw4w9WgXcQ"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Understand This Video For Me AI Plugin
Let any AI watch and hear a video. Give Claude, Cursor, Grok Bot or any MCP client a local video file or a YouTube / Vimeo / Loom link, and it gets back what it can actually read:
8–20 timestamped key frames (scene changes, duplicates removed, small JPEGs)
a timestamped transcript (site captions when available, otherwise local Whisper)
one contact-sheet image with all frames in a grid
one
context.mdtimeline mixing what was shown and what was said
Everything runs 100% on your computer. Your videos never leave it.
Price
$49 one-time. Lifetime licence, up to 3 of your devices. No subscription. Buy: https://africaappstudio-max.github.io/video-context-ai/
No refunds for change of mind. If it won't work on your computer, email me within 7 days and I'll fix it or refund you.
Related MCP server: claude-real-video
Install
Option 1: one click (Claude Desktop). Download understand-this-video.mcpb from Releases and open it. You need ffmpeg installed (Mac: brew install ffmpeg, Windows: winget install ffmpeg).
Option 2: installer (Claude Desktop + Cursor, includes transcription). Download this repo, then:
Mac:
bash install.shWindows: right-click
install.ps1→ Run with PowerShell
The installer adds ffmpeg and Python if missing, installs the plugin, registers it in Claude Desktop and Cursor, and asks for your licence key.
Other MCP clients: add a stdio server with command understand-this-video-mcp (after pip install .), or see .mcp.json.
Activate: understand-this-video activate YOUR-KEY. Internet is needed once to activate; then it works offline.
Example prompts
"Watch ~/Downloads/demo.mp4 and write the step-by-step instructions it shows."
"Summarise this YouTube video with timestamps: https://www.youtube.com/watch?v=..."
"Look at this screen recording of my bug, ~/Desktop/bug.mov, and tell me what went wrong."
Tool
understand_this_video: input source (file path or URL) plus optional options. Returns the frames as images, the transcript and the timeline.
Privacy
Videos are processed locally. The only network call is the licence check with Freemius. See PRIVACY.md.
Support
Demo video used on the website: "The Babbling Brooke" (YouTube zLTDLeDRrw4), CC BY.
Proprietary software © 2026 Africa App Studio. The source is visible, but use needs a paid licence. See LICENSE.
This server cannot be deployed
Maintenance
Related MCP Connectors
Give AI random access to video: timestamped contact sheets + zoom into any start/end range.
Transcripts, summaries, chapters and timestamped answers for video links, for AI agents.
Video analysis AI: transcripts, summaries, visual scenes/shots, clips, answers in natural language.
YouTube transcripts, search, channel/playlist listings and upload tracking for AI agents.
Related MCP Servers
- AlicenseAqualityDmaintenanceLet AI agents watch videos: local transcripts, speaker labels, scenes, chapters and exact-moment search from any video URL or file. Fully local, no API keys.43AGPL 3.0
- AlicenseNot gradedqualityAmaintenanceLets any LLM agent actually watch videos: a watch_video tool takes a URL or local file and returns scene-aware keyframes fused with a timestamped transcript, processed 100% locally with per-source caching.891 PyPI2,234MIT
- AlicenseNot gradedqualityCmaintenanceEnables AI agents to download, convert, and visually analyze videos by providing timestamped frames and locally transcribed speech, all without API keys.MIT
- AlicenseAqualityAmaintenanceEnables AI clients to watch local video files or YouTube/Bilibili and other supported URLs, receiving timestamped transcripts, subtitles, searchable text and keyframe contact sheets. Runs fully offline with local speech recognition and bundled ffmpeg, so no API key or cloud upload is needed.693 PyPI10MIT