OpenScreen 2: Open-Source Screen Recorder and Editor
VideoFreeOpenScreen 2 is an open-source screen recorder and editor for Windows, macOS and Linux. It captures a window or full display, then edits zooms, cursor, captions and framing on a timeline and exports MP4 or GIF.
What OpenScreen 2 Is
The app pairs a screen recorder with a timeline editor, so one capture becomes a finished clip without switching apps. It targets short product demos rather than live streaming and runs on Windows, macOS and Linux.
It continues the original OpenScreen, a Mac-first recorder archived by its creator after version 1.5.0; development moved to the getopenscreen organization under the same name and license.
This is a documented overview, not a hands-on review.
Recording, editing and the demo finish
Choose a window or a whole display. Microphone and system audio are captured with the screen, while the webcam is saved as its own file for later placement. System audio varies by platform: macOS and Windows capture it out of the box, Linux needs PipeWire.
The timeline covers trimming, cropping, speed, annotations, blur regions and transcript-based cuts. The cursor is stored as data, so it can be restyled after recording, and zooms are placed from where you clicked — a rule that needs no model or network call. Recordings open already framed with a background, padding, a shadow and rounded corners.
- Five cursor packs, a 3D cursor, device frames and animated backgrounds
- Export to MP4 at 720p up to source resolution (24, 30 or 60 fps) or GIF, encoded on the GPU
Captions, the AI edit assistant and the command line
Captions are generated on the machine with whisper.cpp. The model downloads once (roughly 264 MB) and then runs offline, so audio is not sent to a caption service. Translating subtitles does send text to the provider you configure.
The AI edit assistant is separate and optional: it stays off until you connect a provider key, then accepts Claude, OpenAI, Gemini, Mistral, OpenRouter, MiniMax or any OpenAI-compatible endpoint and turns chat requests into undoable cuts, zooms, speed changes or camera moves.
A headless command-line tool records and renders with no visible window, with JSON output for scripts and CI — separate from the in-editor assistant.
Platform support, stability and privacy
It supports Windows 10 version 1903 or later, macOS 13 (Ventura) or later, and Linux with PipeWire and xdg-desktop-portal. Older integrated graphics can hit known recording-stop issues.
The repository calls the project actively developed and warns of rough edges and occasional breaking changes to the project file format; test the current release before depending on a fixed workflow. Recording, transcription and rendering stay local; text leaves only when you use the AI chat panel or caption translation, each with a key you supply.
Best For
Recommended use cases and scenarios where OpenScreen 2 shines.
Pros
- Records a window or full display with microphone and system audio, plus a separate webcam file
- Edit on one timeline instead of exporting to a separate video editor
- Cursor and zoom are stored as data, so they can be changed after recording
- Captions run on-device and keep working offline after the model download
- MP4 and GIF export at several resolutions and aspect ratios, with GPU acceleration
- Headless recording and rendering for scripts and CI pipelines
Cons
- Under active development, so features and the project file format can change between releases
- The optional AI edit assistant is off by default and needs a third-party provider key you supply
- Offline captions need a one-time model download before they work without a connection
- The Windows .exe is unsigned and triggers a SmartScreen warning
Frequently Asked Questions
Common questions about OpenScreen 2, answered.
Is OpenScreen 2 free, and can I use it commercially?
Yes. The license allows personal and commercial use, with no watermark, account or usage limit; the only optional cost is a provider key you supply.
Which operating systems does it support?
Windows 10 version 1903 or later, macOS 13 (Ventura) or later, and Linux with PipeWire and xdg-desktop-portal.
Does it need an account or an internet connection?
There is no account. Captions and rendering run locally, and captions work offline after a one-time model download; the optional AI features need a connection.
How does the optional AI edit assistant work?
Turn it on by connecting a key from a supported provider (Claude, OpenAI, Gemini and others). You describe the change you want, and the result lands on the timeline as an ordinary edit you can undo.
What should I know about project maturity?
Treat it as actively developed. The project warns about rough edges and occasional breaking changes, so test the current release before depending on it.
Reviews & Ratings
Loading reviews...
Similar Tools
More Video tools you might like
Loom AI
Async screen recording with built-in AI that writes titles, summaries, chapters, and action items from every recording.
CapCut
CapCut is ByteDance's free AI-powered video editor for short-form content, with auto captions, background removal, and a huge template library for TikTok and Reels.
