Skip to content
imaged
Docs / Setup / Python

Python setup.

pip install the wheel, import imaged, run a family. Five lines to a working detector.

Prerequisites
Python 3.10 or later
Platforms macOS arm64, Linux x86_64, Windows x64
License token JWT issued from /dashboard
  1. 01

    Install

    The platform tarball ships an imaged/ folder with the Python bindings and the right libsdk for your OS. Drop it next to your script, or pip install it locally.

    bash 
    tar -xzf imaged-macos-arm64-0.10.0.tar.gz
pip install ./imaged-sdk-0.10/python
  2. 02

    Hello world

    Construct AI, set a license token from /dashboard, verify, print the version. The Config object is a thin wrapper around the same string-keyed map every binding uses.

    python 
    from imaged import AI

ai = AI()
ai.config.set("license.token", "your-license-token")
ai.config.set("license.audience", "cameras.acme.com")

if not ai.config.check_license() or not ai.licensed:
    raise SystemExit("license check failed")

print(f"imaged {ai.version}")
  3. 03

    Run

    No build step. Plain Python.

    bash 
    python hello.py
# imaged 0.10.0
  4. 04

    Run a family

    detect() returns a list of dicts with label, score, and a pixel-coordinate box. process() mutates an Image in place for upscale, deblur, and rembg.

    python 
    from imaged import AI, Image

ai = AI()
ai.config.set("license.token", "your-license-token")
ai.config.set("license.audience", "cameras.acme.com")
ai.config.check_license()

ai.config.set("yolo.model", "yolo26x.onnx")

img = Image()
img.load("bus.jpg")

for det in ai.detect("yolo", img):
    print(det["label"], det["score"])
  5. 05

    Where to go next

    The detect.py example in the examples repo wires this into OpenCV with a live webcam. Same five SDK calls, plus a frame loop.

Working example
ObjectDetector-python
Where to go next
Setup for another language
C C++ Java Kotlin Objective-C Swift