# System Requirements This section shows the hardware, software, and platform requirements to help you set up and run Smart Traffic Intersection Agent efficiently. The agent currently supports CPU- and GPU-based runs, and runs in the context of video summary pipeline. Hence, the system requirements is as per the documentation in the sample application. Object detection, provided by the Smart Intersection application, can additionally be configured to run on NPU; see [Enabling NPU for Object Detection](../get-started.md#enabling-npu-for-object-detection). ## Supported Operating Systems - Ubuntu 22.04 LTS - Ubuntu 24.04 LTS - Ubuntu 26.04 LTS - Other Linux distributions with Docker support ## Hardware Requirements | Component | Minimum | Recommended | |---|---|---| | CPU | Intel® Core™ i5 or equivalent | Intel® Core™ Ultra Ultra 2 and 3 with integrated GPU or Intel® Xeon® processor | | RAM | 16 GB | 32 GB or more | | Disk Space | 50 GB free | 100 GB free | | GPU (optional) | — | Intel® integrated GPU (iGPU) for accelerated VLM inference | | NPU (optional) | — | Intel® NPU for accelerated object detection | | Network | Internet access for weather API and model downloads | — | ## Software Requirements - Docker Engine version 29.0 or later - Docker Compose v2: [Installation Guide](https://docs.docker.com/compose/install/) - Git (for cloning the repository) ## Other Requirements - **MQTT Broker**: An MQTT broker for traffic data streaming (included in the deployment stack) - **Hugging Face authentication (conditional)**: No Hugging Face account, token, or registration is required to download public, ungated VLM models such as `OpenVINO/Phi-3.5-vision-instruct-int8-ov`. If you choose a gated or private model, you need an account with access to that model and an authorized token. Only for those models, follow the [Hugging Face token instructions](https://huggingface.co/docs/hub/security-tokens) and set `HUGGINGFACE_TOKEN` before running setup. Internet access is still required for model downloads. - **Network Ports**: The agent requires available ports for the backend API (default: 8081) and UI (default: 7860). Ensure these are not in use, or leave them empty in the configuration to use ephemeral ports. ## Validation - Ensure all required software is installed and configured before proceeding to [Get Started](../get-started.md). ## Learn More - [Overview](../index.md) - [API Reference](../api-reference.md)