A free, open-source (MIT) screen recorder and video editor positioned as a Screen Studio alternative. It records on Windows, macOS, and Linux with automatic zooms based on your clicks, smart cursor tracking, webcam overlays (picture-in-picture, side-by-side, full-screen with background removal), and on-device transcription via whisper.cpp (100 languages, offline after one model download). The optional AI editing agent (off by default) lets you describe edits in plain language—\"cut the dead air,\" \"zoom when I open the terminal\"—and executes them as undoable timeline operations, using your own API key from Claude, OpenAI, Gemini, or others. An MCP server lets Claude Code or Codex drive the recording and editing tools directly. No account, no watermark, no subscription, free for commercial use. System requirements: macOS 13+, Windows 1903+ (Intel 8th Gen / Ryzen 2000+), Linux with PipeWire. 8GB RAM minimum, 16GB recommended. Note: the domain is new (July 2026), but the project is MIT-licensed with a public GitHub repo and detailed privacy policy (no servers, local-only recording).

Pros & Cons:

Pros: The free + open-source (MIT) + cross-platform combination is genuinely rare for this category—Screen Studio costs ~$29/mo and is Mac-only, while OpenScreen covers Windows/Linux and costs nothing, with no watermarks or feature gates; the automatic zoom and cursor tracking are core to making product demos look polished without manual keyframing, and the cursor is recorded independently of pixels so you can adjust it after the fact; the on-device transcription (whisper.cpp) enables captions and text-based editing without sending audio to a cloud, preserving privacy; the optional AI agent and MCP support extend it beyond a simple recorder—you can drive edits from Claude Code or Codex, and the AI works through undoable timeline operations rather than re-rendering.

Cons: System requirements exclude older hardware—Windows needs Intel 8th Gen / Ryzen 2000+, macOS needs Ventura+, and 8GB RAM is the floor (16GB recommended), so some users won\'t be able to run it; the AI editing feature requires your own API key and is off by default, so the \"AI agent\" experience isn\'t out-of-the-box; the domain is very new (July 2026) and has low traffic, which triggered a \"high risk\" automated scan on WebSafely despite no actual malware/phishing being detected—users may want to verify the GitHub repo before downloading; Linux audio/capture depends on PipeWire and xdg-desktop-portal, which may not be configured on all distros.

Who is it for?:Developers, product managers, and educators who need to create polished product demos or tutorials and want a free, cross-platform alternative to Screen Studio; privacy-conscious users who want on-device transcription and local-only recording; developers already using Claude Code or Codex who want to automate demo creation via MCP; Windows/Linux users who are excluded by Screen Studio\'s Mac-only limitation. Not suited for users on older hardware (pre-Intel 8th Gen Windows, pre-Ventura macOS), anyone who needs one-click AI editing without configuring an API key, or users who require a long-established domain with extensive third-party security audits.