Let AI agents record the Windows screen, watch it as frames, and edit it with ffmpeg
Screencast MCP is an MCP integration. Let AI agents record the Windows screen, watch it as frames, and edit it with ffmpeg.
The published npm package is @tmhs/screencast-mcp, version 0.52.0. Its declared MCP transport is stdio.
The package declares SCREENCAST_HOME: Where recordings, frames, screenshots, and edits are written (default /.screencast-mcp); SCREENCAST_ALLOWED_DIRS: Semicolon-separated folders tools may read from and write to besides SCREENCAST_HOME; when set (or when the client sends roots), files elsewhere are refused. Also allows overwriting files in these folders; FFMPEG_PATH: Path to ffmpeg if it is not on PATH (ffmpeg and ffprobe are required); FFPROBE_PATH: Path to ffprobe if it is not on PATH; SCREENCAST_WHISPER_MODEL: whisper.cpp model file (ggml-.bin) used by transcribe when a call passes no model; SCREENCAST_MAX_RECORDING_SEC: Default recording cap in seconds when a call sets no maxDurationSec (default 3600; 0 disables it); SCREENCAST_JOB_TIMEOUT_MS: Override the edit job timeout in milliseconds (default scales with the media length); SCREENCAST_RETENTION_DAYS: At startup, delete screenshots, frames, leftover raw event logs, and index_recording indexes older than this many days (never recordings or edits; off by default); SCREENCAST_PS_TIMEOUT_MS: Limit in milliseconds for the PowerShell helpers that list monitors and windows and run OCR (default 20000; OCR 60000); SCREENCAST_CONFIRM_CAPTURE: Confirm each live capture through MCP elicitation: auto (default) asks when the client supports it, 1/always also refuses captures in clients that cannot ask, 0/off never asks; SCREENCAST_TOOLSETS: Comma-separated toolsets to load: capture, watch, edit, produce, audio (doctor and get_media_info are always on; default all); SCREENCAST_DEBUG_JOBS: 1 traces each edit job on stderr (start, a heartbeat every 30 s, and the end) to diagnose a stuck edit; default off.
Check the maintainer’s instructions for account access, authentication and the supported client configuration. Once connected, inspect the tools your client receives and try a small task that matches the server’s stated purpose. The registry description does not include a verified, complete tool list.
View the publisher’s registry entry · Project repository · Project website. This profile uses version 0.52.0 from the registry. Registration identifies the publisher namespace; it does not establish that this is a first-party integration for every service it connects.
Browse related video tools MCP servers for tools that cover the same kind of work.
| Variable | Description | Required |
|---|---|---|
| SCREENCAST_HOME | Where recordings, frames, screenshots, and edits are written (default /.screencast-mcp) | Check documentation |
| SCREENCAST_ALLOWED_DIRS | Semicolon-separated folders tools may read from and write to besides SCREENCAST_HOME; when set (or when the client sends roots), files elsewhere are refused. Also allows overwriting files in these folders | Check documentation |
| FFMPEG_PATH | Path to ffmpeg if it is not on PATH (ffmpeg and ffprobe are required) | Check documentation |
| FFPROBE_PATH | Path to ffprobe if it is not on PATH | Check documentation |
| SCREENCAST_WHISPER_MODEL | whisper.cpp model file (ggml-.bin) used by transcribe when a call passes no model | Check documentation |
| SCREENCAST_MAX_RECORDING_SEC | Default recording cap in seconds when a call sets no maxDurationSec (default 3600; 0 disables it) | Check documentation |
| SCREENCAST_JOB_TIMEOUT_MS | Override the edit job timeout in milliseconds (default scales with the media length) | Check documentation |
| SCREENCAST_RETENTION_DAYS | At startup, delete screenshots, frames, leftover raw event logs, and index_recording indexes older than this many days (never recordings or edits; off by default) | Check documentation |
| SCREENCAST_PS_TIMEOUT_MS | Limit in milliseconds for the PowerShell helpers that list monitors and windows and run OCR (default 20000; OCR 60000) | Check documentation |
| SCREENCAST_CONFIRM_CAPTURE | Confirm each live capture through MCP elicitation: auto (default) asks when the client supports it, 1/always also refuses captures in clients that cannot ask, 0/off never asks | Check documentation |
| SCREENCAST_TOOLSETS | Comma-separated toolsets to load: capture, watch, edit, produce, audio (doctor and get_media_info are always on; default all) | Check documentation |
| SCREENCAST_DEBUG_JOBS | 1 traces each edit job on stderr (start, a heartbeat every 30 s, and the end) to diagnose a stuck edit; default off | Check documentation |
Independent-index web search for assistants — fresh results without the Google API maze.
Manipulates video files using FFmpeg to resize videos and extract audio in various formats.
A Model Context Protocol (MCP) server for YouTube operations, demonstrating core MCP concepts including tools and logging.
Parallel video rendering tools: detect GPU encoders, render, color grade, merge audio, concat.
A server built on the Model Context Protocol (MCP) that enables direct downloading of YouTube video transcripts, supporting AI and video analysis
An MCP server that helps YouTube content creators find perfect background music for their shorts by analyzing script content and recommending tracks