Object Detection with YOLO in 2026

Ultralytics YOLO (60,218 stars, AGPL-3.0) is the easiest object detection library: 10 lines of Python detect 80+ object classes in images, video and live webcam feeds - free and GPU or CPU friendly.

💡 What You Will Learn

Ultralytics YOLO (60,218 stars, AGPL-3.0) is the easiest object detection library: 10 lines of Python detect 80+ object classes in images, video and live webcam feeds - free and GPU or CPU friendly.

📜 Table of Contents

• The short answer
• Detect objects in an image
• Detect in video or webcam
• Beyond detection
• Real numbers
• FAQ

The short answer

Ultralytics YOLO (60,218 stars, AGPL-3.0) bundles the YOLOv8/YOLO11 model family with a dead-simple Python API. Detection, segmentation, classification and pose estimation are each a few lines - no model architecture knowledge required.

Detect objects in an image

pip install ultralytics

from ultralytics import YOLO

model = YOLO("yolo11n.pt")           # nano model: fast, small
results = model("bus.jpg")           # returns detections

for r in results:
    for box in r.boxes:
        print(r.names[int(box.cls)], round(float(box.conf), 2))

Detect in video or webcam

model = YOLO("yolo11s.pt")           # small model: better accuracy
model.predict("video.mp4", save=True)          # annotated video file
model.predict(source="0", show=True)           # live webcam

Beyond detection

Real numbers

• YOLO11n runs at 80+ FPS on a consumer GPU and ~2-5 FPS on CPU.
• The COCO pretrained model detects 80 everyday object classes out of the box.
• Ultralytics is used in production by thousands of companies for counting, inspection and security.

FAQ

Q: Do I need to train it? A: No - the pretrained COCO model works immediately. Train only when you need custom classes (e.g., specific products).

Q: License concerns? A: AGPL-3.0 for the library; Ultralytics also sells enterprise licenses. For commercial closed-source use, check their licensing terms.

Q: Can I export to other formats? A: Yes - model.export(format="onnx") exports to ONNX, TensorRT, CoreML, TFLite and more.

❓ FAQ

Do I need to train it?

No - the pretrained COCO model works immediately. Train only when you need custom classes (e.g., specific products).

License concerns?

AGPL-3.0 for the library; Ultralytics also sells enterprise licenses. For commercial closed-source use, check their licensing terms.

Can I export to other formats?

Yes - `model.export(format="onnx")` exports to ONNX, TensorRT, CoreML, TFLite and more.