Pipeless: Open-Source Computer Vision Framework

Key Features

• Multi-stream Processing

  Supports processing of multiple video streams using protocols such as RTSP, RTMP, HTTP, and files.

• Inference Runtime Support

  Integrates with runtimes including ONNX Runtime, OpenVINO, CoreML, CUDA, and TensorRT.

• Dynamic Stream Management

  Supports adding, editing, or removing streams via a CLI or REST API.

• Multi-language Support

  Supports application code written in Python and Rust.

• Flexible Deployment

  Designed to run on cloud infrastructure, IoT devices, or edge hardware, including environments without an internet connection.

• Code Parallelization

  Handles parallel execution of code to support video analysis in real time.

Use Cases

• Surveillance Object Detection

  Locating and identifying elements within a video stream for security or monitoring purposes.

• Gesture Recognition

  Using pose estimation to identify and track body joints for human-computer interaction.

• Edge AI Deployment

  Deploying computer vision models on IoT or edge devices for local processing.

FAQ

What does Pipeless do?

Pipeless is a framework that handles technical infrastructure for computer vision apps, including memory management, code parallelization, and multimedia pipelines.

Where can Pipeless applications be deployed?

Applications can be deployed to the cloud, IoT devices, or edge devices, and may operate without an internet connection.

How much does the enterprise version of Pipeless cost?

Enterprise pricing starts at $30 per camera or stream monthly.