Open Edge Platform
An Intel-based, partner-enabled, secure, optimized open platform for the edge to accelerate solution development
Modular and Optimized Edge Software Stack
Open Edge Platform 2026.1 is here
Capabilities Highlights
- Evaluate industry-ready multimodal edge AI use cases optimized with blueprints, reference applications, and pre-built pipelines
- Composable, open-source tools, libraries, and microservices for building multimodal vision AI and Gen AI workflows
- AI inferencing at the edge with popular open source models for Vision AI, Gen AI, and Physical AI, optimized on Intel silicon
What's New in 2026.1
AI Suites
- Federal and Aerospace (preview) previews a handheld multimodal AI sample application, with LLM inference, speech-to-text, and chat UI for real-time metrics. New defense, civilian, and aerospace blueprints combine ruggedized mission-computing patterns, spatial intelligence, and sensor-fusion use cases for mission-critical edge AI and control workloads.
- Health and Life Sciences extends multimodal AI patient monitoring sample application to a neonatal ICU warmer workload that detects patient presence, latch safety, caregiver presence and rPPG concurrently detecting conditions and sending real-time alerts when specific conditions and thresholds are met.
- Metro adds curated, use case-based Metro Analytics Catalog of pre-optimized AI models and analytics workflows. New multimodal AI predictive analytics for critical infrastructure use cases, and new Video Surveillance as a Service sample improves camera analytics and connects with commercial VMS solutions. Four transportation workflows added to pipeline zoo for improved developer journey.
- Manufacturing adds new Classifier ML model for weld time series data analysis, multi-camera support for PCB defect detection. Batch processing for user-defined functions enables large-scale image and video processing tasks, complex metrics calculation, custom defect classification, data transformation, VLM fine tuning and explainability. Plus, new framework for concurrent Windows Vision AI pipelines.
- Robotics enables support for LeRobot and LeKiwi robot manipulations and multi-camera GMSL integrations, to optimize devkit hardware performance on Intel® Core™ Ultra Series 3 and OpenVINO™ Physical AI framework.
- Education adds OCR and LLM-based Q&A to audio transcription for precise content search. Now uploads and indexes multimodal assets for live video delivery.
- Retail adds suspicious activity detection pipeline to Loss Prevention sample application for person of interest tracking with real-time streaming protocol (RTSP) . VLM inference is enhanced for running concurrent multimodal AI pipelines, plus new GitHub Actions CI/CD workflows automate deployments and testing integrations.
AI Libraries
- DL Streamer AI Coding Agent (preview) enables vibe coding of hardware-optimized video analytics applications. DeepStream to DL Streamer conversion streamlines license plate recognition use case with a single prompt. Standardizes metadata on Gstreamer GstAnalytics. Pipeline Zoo video analytics pipelines, pre-optimized for Intel GPU and NPU, automate model download and pipeline execution in CLI.
- OpenVINO™ offers prompt lookup decoding for vision-language pipelines and reduces GPU model loading times with cache blobs, preventing bottlenecks in multi-stage agentic AI pipelines.
- OpenVINO™ Physical AI seamlessly integrates cameras, sensors, and Vision-Language-Action models to enables low-latency inference to eliminate motion jitter with tighter control loops, ensuring consistent heterogeneous acceleration on Intel® Core™ Ultra Series 3 processors.
- Physical AI Studio goes from preview to gold with full OpenVINO™ Physical AI framework integration and support for LeRobot, PyTorch, and NVIDIA with pre-packaged containers supporting XPU, CUDA, and CPU environments.
- Geti™ requires fewer CPU resources and memory to fine-tune and run inference directly on edge and client hardware. Adds support for real-time DETR transformer model on DINOv2 for high-quality detection and instance segmentation. Decoupled edge-ready pipelines available for easier integration and faster iteration.
- Geti™ Instant Learn adds Efficient SAM3 model with quantized variants and per-project device selection with hardware-aware optimization. Extends SAM 3 with visual reference sample, plus new text prompting support.
- Scenescape adds support for pose estimation and scene hierarchy to mitigate occlusion in crowded, cluttered scenes, Modernizes embedding vectors and explicit track states for accurate people and object re-identification.
Solve the Biggest Challenges at the Edge
Bring industry-ready edge AI to market sooner, with less risk
Enable industry-ready use cases with pre-built sample applications, benchmarks, and demos to evaluate performance and right-size hardware provisioning to avoid guesswork.
Optimize AI Inference at the Edge
Achieve optimal performance with popular Vision AI and Agentic AI models optimized for Intel silicon with the OpenVINO™ toolkit, a built-in runtime and OpenVINO™ Physical AI and OpenVINO™ Gen AI frameworks, for accelerated AI inference.
Make Multi-modal AI easier to operationalize and scale
Accelerate development cycles from prototyping to production with a curated, open-source collection of tools, pipelines, frameworks, microservices and libraries to enable advanced AI functionality and real-time controls.
Edge AI Suites
Accelerate your solution development and optimization with industry-specific Edge AI Suites.
Retail AI Suite
This suite is an open source application framework designed to accelerate AI workload evaluation and hardware selection for point-of-sale use cases at the edge. This framework helps retailers assess device configurations across Intel product generations to enhance decision-making and reduce the total cost of ownership.
It includes sample applications for use cases such as:
Self Checkout
- Product recognition (detection, classification, and tracking)
- Full pipeline (product, weight, text, and barcode)
- Age verification
Loss Prevention
- Fake scans
- Items in the basket
- Multiproduct identification
- Product switching
- Shopper behavior (hiding items)
- Event video summarizationPerson-of-interest re-identification
- Suspicious activity detection
Order Accuracy
- Object detection using multi-camera computer vision
- Smart video analysis and summarization
- Agentic AI workload
- Advanced LVLM Analysis for Fast Food and Dine-in workflows
Manufacturing AI Suite
This suite is a powerful toolkit for building and scaling AI in industrial environments. The suite simplifies real-time AI development and deployment by using Edge AI technology from Intel.
It supports predictive maintenance, anomaly detection, quality inspection, worker safety, and more. Developers benefit from tools like IoT protocol support, AI analytics libraries, multicamera system software, and closed-loop AI pipelines. Benchmarking resources help evaluate performance across time series, vision, and GenAI use cases.
Sample applications enabled by Manufacturing AI Suite include:
- Predictive maintenance
- Process optimization
- Anomaly detection
- Quality inspection
- Worker safety
- Vision-guided robots
- Operational diagnostics
- Weld defect detection
Metro AI Suite
This suite is an application framework with libraries, tools, and reference implementations, enabling partners to create AI solutions in the video safety and security, transportation, and government edge markets.
Some of the Metro AI Suite components include:
AI Apps for any Intel-Based Edge System
- Advanced video search
- Video summarization
- Natural language search
- Intersection management
- Live video stream analysis
- “Live” avatar interaction
- Smart parking
- Smart tolling
- AI transit route system
- Smart intersection
Benchmarks for Real-World Scenarios
- Benchmarking specifications on Intel® AI Edge Systems
- Platform sizing tool
Tools and Software for AI Acceleration
- AI inference optimization
- Visual and deep learning optimization
- Media processing acceleration
- Edge server video analytics
- Arm* technology to Intel technology conversion
- Sensor fusion-enabled traffic management
Education AI Suite
This preview collection of education-focused AI applications, libraries, and benchmarking tools demonstrate how audio and video recordings from class sessions can be transformed into concise summaries for use in educational settings. Featuring audio-to-speech (ASR) models optimized on OpenVINO™ and streaming media analytics with intelligent pose detection and re-identification (Re-ID) capabilities in DL Streamer, the suite introduces advanced multimodal pipelines to demonstrate high-performance deployment on Intel® CPUs, integrated GPUs, and NPUs.
Audio Intelligence
- Audio transcription with ASR models (Whisper, Paraformer)
- Summarization with LLMs (Qwen, LLaMA)
- Plug-and-play architecture for integrating new ASR and LLM models
- API-first, design-ready, front-end integration
- Extensible roadmap for real-time streaming, diarization, translation, and video analysis
Video Intelligence
- Front Camera Pipeline: Student pose detection: sitting, standing, hand raise, leaning
- Rear Camera Pipeline: Re-Identification (ReID) to track students across camera views
- Board Camera Pipeline: Board content Classifications
Robotics AI Suite
Accelerate development adapting physical AI at the edge and explore new functionality with modular, open-source software components to enable fast development of humanoid, autonomous mobile, and stationary robotics. Sample applications with ROS 2 support, demonstrate vision AI, Gen AI, and media analytics pipelines for enabling robotic manipulation, perception, locomotion, and imitation learning with real-time controls. AI inference engines, and hardware-aware tuning enable fast deployment on Intel silicon.
Collection reference applications, libraries, and pipelines for:
Humanoid
- Imitation Learning – ACT
- LLM Robotics Demo
- ROS 2 and Open Standards
- AI Inference engine
- Hardware-aware tuning
- VSLAM: ORB-SLAM3
Stationary
- Vision & Control
Autonomous Mobile
- ADBScan
- Collaborative SLAM
- Segmentation
- Simulations
- Wandering
- ITFastmapping
- GroundFloor S-Planner
- Multi-camera Demo
- Object Detection
Health and Life Science AI Suite
Combine concurrent tracking of real-time vitals, signal processing into vision AI pipelines. Real-world patient monitoring workloads simulate vitals, track activities, detect health changes with rPPG, and send alerts with validated reference pipelines, using custom datasets and Get-fine-tuned models.
Reference Applications
- Multi‑parameter Monitor: Simulate, analyze, and play back metrics and waveforms
- 3D Visual Tracking: Contactless patient presence and activity monitoring
- AI‑ECG Arrhythmia: Signal‑processing with applied AI
- rPPG: Camera‑based heart and respiration monitoring
- Multimodal Patient Monitoring
- NICU Warmer Monitoring Application
Models
- Multi-Task Temporal Shift Convolutional Attention Network (MTTS-CAN) model
- Multimodal patient monitoring Data and Control Flows – Metrics Collector Service
Federal and Aerospace (Preview)
Use capabilities within the Federal and Aerospace AI suite for developing solutions that bring AI to mission –critical use cases including application blueprints to jump start the evaluation of Intel silicon and streamline development of your own custom solutions based on vision AI, language models, and other machine learning tools optimized for Intel hardware. Build with AI libraries tailored to help adapt multi-camera, multi-scene video analytics applications performantly on Intel hardware to improve cost-efficiency and TCO.
Technology and Capabilities
- Multi-modal AI reference
- AI Inference Optimization
- Visual & Deep Learning Optimization
- Media Processing Acceleration
- CUDA to OpenVINO™ compatibity
- Sensor Fusion-Enabled Air Traffic Control
Blueprints
- Enhanced situational awareness for soldier systems and emergency management
- Drone based aerial digital twins powered by Scenescape
Edge AI Libraries
Composable ingredients for building production-quality multimodal edge AI apps, along with user-friendly workflows for optimizing and deploying models.
- Build edge AI applications and use cases that include vision AI, generative AI (GenAI), and time series AI capabilities.
- Get reusable building blocks, with benchmarks, optimized to run performantly on Intel hardware.
- Shorten app development time and simplify model fine-tuning, retraining, and deployment workflows.
Resources
Join the Ecosystem: Develop and Contribute to Open Edge Platform
Discover how the Edge AI Portfolio can help you achieve your business goals.
Provide configured and benchmarked AI hardware to help accelerate your time to market for in-demand AI solutions.
Stay Up to Date
Learn about everything new with Open Edge platform, by signing up you get early access to you get early access to product updates and releases, exclusive invitations to webinars and events, training and tutorial resources and other news.