Skip to content


Switch branches/tags


Failed to load latest commit information.
Latest commit message
Commit time
Aug 24, 2021

DepthAI API Demo Program

Discord Forum Docs License: MIT

This repo contains demo application, which can load different networks, create pipelines, record video, etc.

Click on the GIF below to see a full example run

depthai demo

Documentation is available at


There are two installation steps that need to be performed to make sure the demo works:

  • One-time installation that will add all necessary packages to your OS.

    $ sudo curl -fL | bash

    Please follow this installation page to see instructions for other platforms

  • Python dependencies installation that makes sure your Python interpreter has all required packages installed. This script is safe to be run multiple times and should be ran after every demo update

    $ python3


This repository and the demo script itself can be used in various independent cases:

QT GUI usage

See instuctions here

qt demo

Command line usage


python3 -gt cv - RGB & CNN inference example

python3 -gt cv -vid <path_to_video_or_yt_link> - CNN inference on video example

python3 -gt cv -cnn person-detection-retail-0013 - Runs person-detection-retail-0013 model from resources/nn directory

python3 -gt cv -cnn tiny-yolo-v3 -sh 8 - Run tiny-yolo-v3 model from resources/nn directory and compile for 8 shaves


cv demo

For the full reference, run $ --help.

DepthAI Apps

We currently have 2 apps, uvc and record.

UVC app

This app will upload an UVC pipeline to the connected OAK camera, so you will be able to use an OAK as a webcam.

Record App

Record app lets you record encoded and synced video streams (MJPEG/H265) across all available devices into .mp4, Foxglove's .MCAP, or ROS' .bag . Since mono streams are synced, you will be able to reconstruct the whole stereo depth perception.

Run using $ -app record [-p SAVE_PATH] [-q QUALITY] [--save STREAMS] [-fc] [-tl]. More information about the arguments and replaying can be found here.

Supported models

We have added support for a number of different AI models that work (decoding and visualization) out-of-the-box with the demo. You can specify which model to run with -cnn argument, as shown above. Currently supported models:

- deeplabv3p_person
- face-detection-adas-0001
- face-detection-retail-0004
- human-pose-estimation-0001
- mobilenet-ssd
- openpose2
- pedestrian-detection-adas-0002
- person-detection-retail-0013
- person-vehicle-bike-detection-crossroad-1016
- road-segmentation-adas-0001
- tiny-yolo-v3
- vehicle-detection-adas-0002
- vehicle-license-plate-detection-barrier-0106
- yolo-v3

If you would like to use a custom AI model, see documentation here.

Usage statistics

By default, the demo script will collect anonymous usage statistics during runtime. These include:

  • Device-specific information (like mxid, connected cameras, device state and connection type)
  • Environment-specific information (like OS type, python version, package versions)

We gather this data to better understand what environemnts are our users using, as well as assist better in support questions.

All of the data we collect is anonymous and you can disable it at any time. To do so, click on the "Misc" tab and disable sending the statistics or create a .consent file in repository root with the following content

{"statistics": false}

Reporting issues

We are actively developing the DepthAI framework, and it's crucial for us to know what kind of problems you are facing. If you run into a problem, please follow the steps below and email

  1. Run and share the output from (log_system_information.txt).
  2. Take a photo of a device you are using (or provide us a device model)
  3. Describe the expected results;
  4. Describe the actual running results (what you see after started your script with DepthAI)
  5. How you are using the DepthAI python API (code snippet, for example)
  6. Console output