'%20/%3e%3c/svg%3e)
VisionBench
Everything you need to | your computer vision pipelines.
It's instant.
VisionBench is a versatile desktop tool built for developers and robotics teams. Born from the world of competitive robotics, it provides a native OpenCV environment, letting you rapidly test and refine computer vision pipelines right on your computer.
Zero Hardware Required
Test with webcams, static images, or pre-recorded videos—no robot needed.
Instant Feedback
Stop wasting time on the slow build-and-deploy cycle. Test instantly, deploy once.
Supercharge your development workflow.
From live tuning to node-based editing.
Live Pipeline Testing
Run your pipelines in a simple UI with live feedback from a webcam, image, or video file.
FTC SDK Simulation
Mimics the FTC SDK structure, allowing for easy copy-pasting of code to your robot project.
Real-Time Variable Tuning
Adjust pipeline variables on the fly and see their effect instantly without recompiling.
Integrated VisionGraph Node Editor
Prototype algorithms visually with a node editor. Included in VisionBench.
Multiple Input Sources
Test with webcams, static images, or pre-recorded video files to simulate any condition.
AprilTag Support
Includes full support for the modern `VisionPortal` API and AprilTag detection.
Get VisionBench
VisionBench relies on OpenCV's native libraries, making it performant but platform-specific. Check below to ensure your system is supported.
Download InstructionsWindows
- Intel & AMD (64-bit) Tested
- Older PCs (32-bit) Untested
macOS
- Intel Macs Tested
- Apple Silicon (M-Series) Unofficial*
Linux
- Intel & AMD (64-bit) Tested
- ARM64 (RPI) Unofficial*
*Both Apple Silicon (macOS) and Linux ARM (Raspberry Pi) are not actively tested when new features are developed; however, they are known to work correctly.