Use YOLOv8 with a Webcam Stream

Many computer vision models are deployed using a webcam as an input. The Roboflow Inference Python package enables you to access a webcam and start running inference with a


model in a few lines of code.

In this guide, we will show you how to run


on frames from a webcam stream.

We will:

1. Install supervision and Inference
2. Load the webcam stream and define an inference callback
3. Test the webcam stream

Without further ado, let's get started!

Step #1: Install Dependencies

For this tutorial, we will be using supervision, Inference, and OpenCV. supervision provides a range of utilities you can use in computer vision projects. Inference provides a concise utility through which we can load webcam streams. OpenCV is helpful for annotating video frames when supervision does not have a tool we can use.

Run the following command to install the dependencies we will use in this guide:

pip install supervision inference opencv-python

Once you have installed supervision and Inference, you are ready to define an inference callback and configure a webcam stream.

Step #2: Define an Inference Callback

The Inference Python package provides an inference.Stream() method through which you can access a webcam and run logic on each frame. This callback method works with webcam and RTSP streams, but for this guide we will focus on webcam streams.

In the code below, we use inference.Stream() to read webcam frames and then run inference on each frame using a



import cv2
import inference
import supervision as sv

annotator = sv.BoxAnnotator()

def render(predictions, image):
    classes = {item["class_id"]: item["class"] for item in predictions["predictions"]}

    detections = sv.Detections.from_roboflow(predictions)


    image = annotator.annotate(
        scene=image, detections=detections, labels=[classes[i] for i in detections.class_id]

    cv2.imshow("Prediction", image)


Above, replace "microsoft-coco/9" with the model ID of a YOLOv8 model hosted on Roboflow.

If you have trained a YOLOv5 and YOLOv8 detection, classification, or segmentation model, or a YOLOv7 segmentation model, you can upload your model to Roboflow for use in running inference on your RTSP video stream.

To upload a model to Roboflow, first install the Roboflow Python package:

pip install roboflow

Then, create a new Python file and paste in the following code:

from roboflow import Roboflow

rf = Roboflow(api_key="API_KEY")
project = rf.workspace().project("PROJECT_ID")
project.version(DATASET_VERSION).deploy(model_type="yolov8", model_path=f"{HOME}/runs/detect/train/")

In the code above, add your API key and the path to the model weights you want to upload. Learn how to retrieve your API key. Your weights will be uploaded to Roboflow. Your model will shortly be accessible over an API, and available for use in Inference. To learn more about uploading model weights to Roboflow, check out our full guide to uploading weights to Roboflow.

Above, we:

1. Import the required dependencies.
2. Define the model we want to use.
3. Define a callback function called render() which takes in the predictions from a model and a frame and processes them. We have included some example code to show how to annotate predictions and display them on camera for use in your code.
4. Use inference.Stream() to access a webcam and run our model.

The render() function is run on each frame retrieved from the webcam stream.

In the code above, replace the API_KEY value with your Roboflow API key. Learn how to retrieve your Roboflow API key.

Step #3: Test the Webcam

Now that you have configured a model and webcam stream, you are ready to test your webcam.

Next steps

supervision provides an extensive range of functionalities for working with computer vision models. With supervision, you can:

1. Process and filter detections and segmentation masks from a range of popular models (YOLOv5, Ultralytics YOLOv8, MMDetection, and more).
2. Display predictions (i.e. bounding boxes, segmentation masks).
3. Annotate images (i.e. trace predictions, draw heatmaps).
4. Compute confusion matrices.

And more! To learn about the full range of functionality in supervision, check out the supervision documentation.

Learn how to use other models with a webcam stream

Below, you can find our guides on how to use more computer vision models with a webcam stream.