Screencap

Screencap is a macOS application that records screen activity, clicks, keystrokes, and window context, then uses an on-device AI model to automatically segment the footage into labeled tasks with speech and visual indexing, making past work searchable. The goal is to turn how a team actually performs its work into structured data that can be shared, used for training new employees, or used to train computer-use AI agents. Privacy is built into the recording pipeline: sensitive applications like password managers and banking apps are excluded from capture before recording starts, and content such as email or chat is automatically masked. Screencap also integrates with MCP servers such as Gusto, Attio, Linear, and Notion to pull in contextual information about what a recorded task relates to. Users who want to contribute to AI research can optionally donate anonymized recordings to help train computer-use models, and the application's code is source-available on GitHub under a PolyForm noncommercial license. Pricing starts at $9/month for local-only use on a single Mac, with a $20/month tier adding personal cloud storage and custom pricing for team cloud plans; it's free for personal, research, and nonprofit use. The tool currently only runs on Apple silicon Macs, with Windows, a browser extension, and Android listed as coming soon.