• 8 min read
OCR Lot Code and Expiry Date Verification for Medical Packaging
Train a Roboflow localization model, isolate printed label fields, and verify batch numbers and expiry dates with Google Gemini.
• 2 min read
Launch: Use YOLO26 Semantic Segmentation with Roboflow
We are excited to announce support for YOLO26 semantic segmentation in Roboflow.
• 8 min read
Automate Medical Device Packaging Inspection and UDI Label Verification
Use RF-DETR and Gemini 2.5 Pro to inspect medical device packaging, verify UDI labels, and identify packaging issues.
• 2 min read
Claude Fable 5 for Vision: Evaluation and Benchmarks
Claude Fable 5 is a strong reasoning model for visual understand but not a state-of-the-art vision model.
• 12 min read
Give My Agent Eyes
Combine RF-DETR and LMMs to build an AI pipeline that perceives, reasons, and acts.
• 5 min read
What Goes in a Vision AI Solution Catalog
Scale your enterprise computer vision with a Standard Solution Catalog. Learn the six artifacts needed to move from isolated pilots to global deployments.
• 4 min read
What Is YOLO27?
YOLO27 brings 3D perception to YOLO with YOLO-Depth and YOLO-StereoDepth.
• 7 min read
Turn Vision AI Detections Into Alerts: PLC, MES, Slack
Learn how to detect defects with Roboflow and trigger real-time downstream alerts in manufacturing facilities.
• 12 min read
YOLO vs. VLMs: When to Use Each
YOLO and VLMs are not competing technologies. YOLO is good for speed, edge deployment, and fixed-class precision. VLMs good on reasoning, flexibility, and open-vocabulary tasks. Learn which one fits your use case and how to use it in Roboflow Workflows.
• 4 min read
How to Build a Vision AI Center of Excellence
Build a vision AI Center of Excellence in three steps: define a North Star metric, pilot your easiest use case, codify the first catalog entry.
• 5 min read
Who Runs a Vision AI Program: The Builder, Scaler, Operator Model
A vision AI team needs three roles: Builder, Scaler, and Operator. Here is what each one owns and how to staff them as your program scales.
• 5 min read
Why Top-Down Vision AI Rollouts Fail: The Push Trap vs. the Pull Model
Top-down vision AI rollouts stall because plants reject what they didn't specify. Here's why the Push Trap fails and how the Pull Model scales.
• 14 min read
Building a Real-Time Snack Detection Web App
Learn how to build a real-time computer vision web app. Combine Roboflow object detection, MediaPipe, and Flask to build a custom snack-tracking app.
• 4 min read
Cosmos 3: Evaluation for Vision Use Cases
NVIDIA Cosmos 3 offers zero-shot performance on fixed-camera footage across airports, warehouses, and production lines.
• 7 min read
Is Your Vision AI Program in Pilot Purgatory? A 10-Question Diagnostic
Stuck running vision AI pilots that never scale? Answer these 10 yes/no questions to score whether your program is in pilot purgatory or on the curve.
• 5 min read
Renting vs. Owning Vision AI: How to Pick a Vendor
The right vision AI vendor makes you need them less over time. Here's how to tell renting from owning, and the questions to ask before you sign.
• 7 min read
AI Cameras vs. IP Cameras: What's the Difference?
A practical guide to understanding how AI cameras and IP cameras differ, and when each makes sense for your operation.
• 9 min read
Sports Analytics AI with Roboflow
Learn how to build a sports analytics AI pipeline. Use RF-DETR and Gemini to automate sports formation analysis and player tracking.
• 10 min read
YOLO for Anomaly Detection
Learn how to build a YOLO based supervised anomaly detection system with Roboflow. This tutorial shows how to collect normal and defective alloy wheel images, label defects with bounding boxes, train a YOLO11 model, and deploy it to detect scratches, and paint damage.
• 4 min read
Roboflow Advances Defect Detection Pipeline for Manufacturers with Synthetic Data Powered by NVIDIA Physical AI
SUMMARY
Roboflow and NVIDIA have integrated synthetic data generation into the Roboflow platform for manufacturing defect detection, combining the NVIDIA Defect Image Generation skill, the Cosmos world foundation model, and NVIDIA Isaac Sim to generate photorealistic, physics-consistent defect images from a small set of real examples. In a benchmark
• 5 min read
What Is YOLO-StereoDepth?
YOLO-StereoDepth brings stereo depth for robotics to the YOLO family in September 2026. See what we know and how to read metric depth today.
• 4 min read
What Is YOLO-Depth?
YOLO-Depth brings monocular depth estimation to the YOLO family in September 2026. See what we know and how to estimate depth today.
• 10 min read
Process RTSP Streams for Real-Time Video Analytics
Ingest RTSP streams, handle frame buffering, and run wildfire smoke detection on the Roboflow Inference Docker container.
• 7 min read
Chaining Models: Combining Detection, OCR, and an LLM in a Single Workflow
Learn how to build a multi-stage computer vision pipeline. Chain RF-DETR object detection, OpenAI OCR and LLMs in Roboflow to process documents.
Stay Connected