abhitronix/vidgear

VidGear is a Python framework for working with video in real time. It wraps around well-known libraries like OpenCV, FFmpeg, and ZeroMQ, giving developers a single consistent interface instead of forcing them to learn the quirks of each underlying tool separately. The stated goal is letting you write less code to accomplish more. The framework is organized around a set of components it calls Gears, each focused on a specific job. CamGear reads video from IP cameras, USB cameras, network streams, and streaming site URLs. PiGear targets the camera modules on Raspberry Pi single-board computers. ScreenGear captures whatever is currently displayed on your screen. VideoGear is a general-purpose capture API that also includes a built-in video stabilizer. WriteGear encodes and compresses video to files or streams. StreamGear transcodes video into adaptive streaming formats like HLS and DASH, which are the formats used by most video platforms to adjust quality based on connection speed. NetGear transfers video frames between machines over a network. WebGear and WebGear_RTC serve live video to web browsers, with WebGear_RTC using the browser-native WebRTC protocol for direct peer-to-peer delivery. All of these components run with multi-threading or asyncio, meaning they are designed to keep video flowing without blocking your program. The README describes this as silently delivering robust error-handling and real-time performance, though the exact error handling behavior is documented separately on the project site. Installation is via pip, and the project provides Docker examples for containerized deployments. The README links to documentation covering each Gear separately and includes a guide for developers already familiar with OpenCV who want to switch over. The source code is licensed under Apache 2.0. The full README is longer than what was shown.