# media-context-mcp [Health: Active]

**Category:** 🎥 Multimedia Process  
**Repository:** https://github.com/vishalguptax/media-context-mcp  
**GitHub Stars:** 2  
**npm Downloads (last month):** 252  
**Views:** 3  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/media-context-mcp

## Description
Local MCP server to analyze video, audio & images: frames, transcripts, on-screen text.

## Tools
Capabilities this server exposes over MCP:

- **check_media_deps** — Report which external binaries (ffmpeg, ffprobe, yt-dlp, whisper, tesseract) are available. Call this first if analyze_media fails with a missing-binary error.
- **analyze_media** — Turn a local media file or URL (video, audio, or image) into compact context a model can read — fully local, no paid APIs. Video: montage frames (mode 'sheet', cheapest default), individual stills ('frames'), or scene changes ('scenes'); add transcript and/or ocr. Audio: speech transcript. Image: the picture plus optional OCR. For app/screen recordings use detail:'high' + ocr:true. To catch a transient UI glitch (a flicker/jump lasting <1s), use mode:'filmstrip' with a narrow startSec/endSec window, a high fps (10–15), and a crop around the affected control — it stacks dense frames so you can spot a frame whose value disagrees with the visual. Use the cheap default for everything else. Pass context to frame the analysis.

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

```json
"mcpServers": {
  "media-context-mcp": {
    "command": "npx",
    "args": ["-y","media-context-mcp"]
  }
}
```

## Documentation

## What media-context-mcp MCP server does

The media-context-mcp MCP server turns video, audio, and image sources into context that an AI assistant can use. Sources can be local file paths or HTTP(S) URLs. For video, it can return sampled frames, full-size stills, scene-change frames, or dense filmstrips. Audio can be converted to a transcript, and images or video frames can be processed for text displayed on screen.

The server also removes temporary data after each tool call. Supported link sources include YouTube, Vimeo, and more than 1,000 other sites, according to the project documentation.

## How it works

The primary `analyze_media` tool detects whether the source is audio, video, or an image. Video analysis defaults to `sheet` mode, which places sampled frames into one or two contact sheets. `frames` returns individual stills, `scenes` focuses on scene changes, and `filmstrip` samples densely for brief interface glitches or other short-lived events.

Analysis can be narrowed to a time range or a cropped region. Options also control image format, frame count, frame scale, grid size, scene sensitivity, and sampling rate. Transcription can be enabled for video and audio, with Whisper model choices from `tiny` through `large`. OCR can extract visible text and accepts language and page-segmentation settings.

## Setup and configuration

Install the MCP package in a compatible client with the command `npx -y media-context-mcp`. The README provides configurations for Claude Desktop, Claude Code, Cursor, VS Code, Windsurf, Cline, Roo Code, Gemini CLI, Zed, Codex CLI, and JetBrains AI Assistant. Claude Code can also install it as a plugin, and project-scoped configuration is supported there.

After adding the server, run `npx media-context-mcp setup` for the core helpers. Add `--audio` to enable transcription support. The setup process uses the operating system package manager, and the server discovers the installed helpers automatically. Run `check_media_deps` to inspect available capabilities; `setup --uninstall` removes the helpers.

## Tools and capabilities

The media-context-mcp MCP server exposes two documented tools:

- `analyze_media` accepts a local path or URL and returns model-readable media context. It supports cropping, time windows, language selection, sampling controls, transcription, and OCR.
- `check_media_deps` reports which local media capabilities are ready.

Useful workflows include summarizing a video from keyframes, reading an error from a screenshot, transcribing a meeting recording, explaining a screen recording from scene changes, or locating a brief visual defect with filmstrip sampling.

## Limitations and notes

Transcription depends on the optional audio setup and a locally available Whisper model. OCR language support depends on the language codes and local helper availability. More detailed modes and parameters may increase processing time or the amount of visual context returned.

_Full upstream README: https://allmcps.com/mcp/media-context-mcp/readme_