roboflowroboflowOfficialยท1 Agent Skills Included

inference

Self-host computer vision models and video AI workflows anywhere

Runs computer vision models locally or on edge devices and exposes them through an HTTP API, CLI, and Python SDK. Chains models into visual workflows to detect, track, count, and measure objects in images and live video streams. Removes the need for manual model serving setup with ready-made Docker images for CPU, GPU, and Jetson devices.
npx skills add roboflow/inference --all -g -y
Available:

Gives AI agents the repository layout, environment variables, build commands, and test instructions needed to work on the inference server, CLI, and SDK packages correctly.

All Skills in This Repository (1)

Pure Emerald Level Indicators

Frequently Asked Questions

FAQPage Schema
How to install Roboflow Inference?โ–ผ

Run `npx skills add roboflow/inference --all -g -y` in your terminal to install all skills in this suite globally.

What does Roboflow Inference do?โ–ผ

It turns any computer or edge device into a computer vision server that runs detection, segmentation, and classification models through a local API. You can also chain models into workflows that track, count, and analyze objects in video streams.

Can I run vision models without a GPU?โ–ผ

Yes. The project ships pre-built Docker images for CPU, GPU, and NVIDIA Jetson devices, so you can run models on almost any hardware.

How do I process live camera or RTSP video streams?โ–ผ

The inference server connects directly to webcams and RTSP streams and runs your chosen workflow on each frame with hardware acceleration handled automatically.

Do I need to write code to build a vision workflow?โ–ผ

No. Workflows are composable blocks you can arrange in a visual UI, and the Python SDK lets developers trigger them with a few lines of code.

Related Repositories in Software Engineering

View All in Software Engineeringโ†’