Log inSubmit server
Openscreen logo

Openscreen

by getopenscreen

Drive OpenScreen over stdio to record, edit, caption, annotate, and export screen-capture videos.

InstallRepositoryWebsite
Full README Compare alternatives Embed widget Claim this listing

Install Openscreen

Pick your client, then add it in one click or paste the config.

No confirmed install command yet

We only publish a config when the install details come from the project itself — its README, its docs, or a verified owner. We haven’t found those for getopenscreen/openscreen, and we’d rather show nothing than a guess you’d paste into your client.

Open setup steps on GitHub

Overview

getopenscreen/openscreen connects an AI agent to the OpenScreen desktop app through its MCP mode over stdio. Agents can record screens, edit projects with trims, zooms, captions, and annotations, then export MP4 or GIF videos. It also supports custom WGSL shader effects that render as previewed. Reach for it when an agent needs to produce product demos, walkthroughs, or change videos rather than only manipulate raw media files.

Use cases

  • Record a product walkthrough from an agent-driven desktop session
  • Add a demo video for a code change to a pull request
  • Trim and annotate screen recordings for documentation
  • Export launch or feature videos as MP4 or GIF

Key features

  • stdio MCP control through the OpenScreen executable
  • Window and full-screen recording
  • Timeline editing with zooms, trims, captions, and annotations
  • On-device caption transcription with editable transcripts
  • MP4 and GIF export with GPU encoding and CPU fallback
  • Headless recording and export through the CLI

Security & health

What it can reach and how reliably it runs — from our own checks, not the maintainer’s claims.

ActiveLast checked Today

Recent health check succeeded.

Adoption & maintenance

Factual signals from GitHub, npm, and our automated checks — not a rating.

GitHub stars5kStargazers on the source repository.
Last commitTodayMost recent push to the default branch.

How Openscreen works

What getopenscreen/openscreen does

getopenscreen/openscreen exposes OpenScreen's desktop recording and editing workflow to AI agents. OpenScreen captures a selected window or the full screen, along with microphone and system audio, and can add a webcam picture-in-picture overlay. Agents can then work with the resulting project to create a finished demo or walkthrough.

Editing features include trimming, cropping, segment-level speed changes, cursor styling, automatic or manually placed zooms, captions, text, arrows, images, camera framing, and other timeline changes. Captions are transcribed on the device and the transcript can be edited to remove corresponding parts of the video. Custom WGSL shader effects are also supported, with exports matching the preview.

The finished result can be rendered as MP4 or GIF in multiple aspect ratios and resolutions. OpenScreen's export pipeline uses GPU encoding where available, with a CPU fallback.

How it works

The MCP connection starts the OpenScreen executable with the mcp argument. The desktop application provides the recording and editing capabilities while the agent sends requests over standard input and reads responses over standard output.

OpenScreen also supplies a headless command-line interface for scripts, CI jobs, and coding agents. Projects use the .openscreen format, which can be edited programmatically as JSON. The CLI can record for a specified duration and export a project without opening visible windows; JSON output is available with --json.

For larger workflows, the OpenScreen demo skill helps an agent set up the application and MCP connection, plan a storyboard, record an application, edit the take, export the video, and inspect the result. The workflow can be used for videos attached to pull requests or for launch demonstrations, with publication requiring user approval.

Setup and configuration

Install OpenScreen using the recommended installer for the target platform. macOS builds are distributed as a DMG, Windows uses the Microsoft Store route, and other platform installers are available from GitHub Releases. The application supports macOS, Windows, and Linux.

In Claude Code, add the executable as an MCP server:

Terminal
claude mcp add openscreen -- "<executable>" mcp

The demo skill can be installed with:

Terminal
npx skills add getopenscreen/openscreen --skill openscreen-demo

On macOS, OpenScreen requires Screen Recording and Accessibility permissions. Linux native capture and system audio require xdg-desktop-portal and PipeWire; browser capture remains available with fewer capabilities when those components are absent. The documented minimum RAM is 8 GB, with 16 GB recommended.

Tools and capabilities

  • Record a specific window or the entire screen.
  • Capture microphone and system audio.
  • Add and configure webcam overlays.
  • Apply cursor effects, smoothing, zooms, and camera framing.
  • Generate, edit, and optionally translate captions.
  • Trim, crop, annotate, and adjust playback speed.
  • Apply custom WGSL shader effects.
  • Export projects to MP4 or GIF.
  • Operate headlessly through the CLI with NDJSON output.

Read the full README View source on GitHub

Reviews

No reviews yet — be the first to share how it worked for you.

Frequently Asked Questions about Openscreen

getopenscreen/openscreen is an MCP server that connects AI agents to the OpenScreen desktop screen recorder and video editor over stdio. Its main tools support recording windows or screens, editing projects with trims, zooms, captions, and annotations, applying shader effects, and exporting finished videos as MP4 or GIF files.

Install the OpenScreen desktop app, then configure the app's executable with the `mcp` argument. In Claude Code, run `claude mcp add openscreen -- "<executable>" mcp`. The optional demo workflow is installed with `npx skills add getopenscreen/openscreen --skill openscreen-demo`.

Yes, getopenscreen/openscreen is described for use with Cursor as well as Claude Code and Codex. Configure the OpenScreen executable as an MCP server in the client, using the executable's `mcp` argument to start the stdio connection.

Yes, OpenScreen provides a headless CLI for recording and exporting without visible windows. Use commands such as `openscreen record` and `openscreen export`, and add `--json` to receive NDJSON output for scripts, CI, and coding-agent workflows.

Related MCP servers

All in Multimedia Process All alternatives
Add a badge to your READMEShow it’s listed — live SVG, Markdown or HTML.

AllMCPs Directory Badge

Full Badge Customizer

Showcase your server listing on GitHub or your project documentation. Embed this dynamic SVG badge to highlight official listing status and live engagement.

Badge Style:
Live Dynamic SVG PreviewOpenscreen AllMCPs Directory Badge
Markdown (GitHub README)
[![AllMCPs](https://allmcps.com/api/badge/getopenscreen-openscreen?style=directory)](https://allmcps.com/mcp/getopenscreen-openscreen)
HTML Embed
<a href="https://allmcps.com/mcp/getopenscreen-openscreen"><img src="https://allmcps.com/api/badge/getopenscreen-openscreen?style=directory" alt="Openscreen on AllMCPs" /></a>