Nft-Marker-Creator-App

This editor creates NFT markers for WebARKitLib and ARTOOLKIT 5.x, they are compatible with jsartoolkitNFT.js, jsartoolkit5.js, artoolkit5.js, ARnft.js and AR.js. A Node app and a native Python package (on PyPI) are provided — jump to the Python package section for pip install nft-marker-creator.

This project is based on the original NFT Marker Creator by Carnaux but has been updated to work with the latest versions of Node and NPM, and also to work with the latest version of Emscripten, plus other improvements.

Now with threading support!! Read more in Run the app section

Quick start ⚡

cd src
node NFTMarkerCreator.js -i PATH/TO/IMAGE --threaded <number_of_threads>

Example:

cd src
node NFTMarkerCreator.js -i pinball.jpg --threaded 8

Outputs land in src/output. See the sections below for setup, flags, and optional demo steps.

Prerequisites ✅

• Tested on Windows 11 (PowerShell 5.1) and Ubuntu 22.04 LTS (bash)
• Git 2.39+
• Node.js v22.21.1 (.nvmrc, Node 22 LTS; minimum supported >=20.17.0) and npm 10+
• nvm 0.39+ (optional but recommended for managing Node versions)
• Docker 24+ (required only for the containerised build flow)

Node app

Installation 📦

1. Clone this repository (or download the latest prebuilt NFTMarkerCreator.js and build/ bundle from the GitHub Releases page).
2. Install Node via nvm (recommended):
   • nvm install
   • nvm use 22
3. Install dependencies: npm install

Run the app 🚀

1. Place the image you want to convert inside the src folder (either directly or inside a subfolder).
2. Generate the marker:

 cd src
 node NFTMarkerCreator.js -i PATH/TO/IMAGE

Example:

cd src
node NFTMarkerCreator.js -i pinball.jpg

The output files are saved in the src/output directory. Use --threaded <count> to speed up processing; start with the number of physical CPU cores and adjust based on thermals and available memory (e.g., laptops may peak at 4 threads, workstations can push higher):

cd src
node NFTMarkerCreator.js -i PATH/TO/IMAGE --threaded <number_of_threads>

Example:

cd src
node NFTMarkerCreator.js -i pinball.jpg --threaded 4

CLI flags ⚙️

Example: node NFTMarkerCreator.js -i image.png -level=4 -min_thresh=8

Flag	Description	Default
-zft	Create only the .zft file	—
-noConf	Skip the confidence confirmation prompt	—
-Demo	Generate the demo configuration alongside the marker	—
-level=n	Feature density preset from 0 (few) to 4 (many)	2
-sd_thresh=<value>	Override standard deviation threshold	Auto
-max_thresh=<value>	Override maximum feature threshold	Auto
-min_thresh=<value>	Override minimum feature threshold	Auto
-leveli=n	Initial feature density preset from 0 to 3	1
-feature_density=<value>	Manual feature density multiplier	Auto
-dpi=f	Force a DPI value instead of the embedded JPEG DPI	Image metadata
-max_dpi=<value>	Cap the maximum DPI processed for multi-scale markers	Auto
-min_dpi=<value>	Floor the minimum DPI processed for multi-scale markers	Auto
-background	Run detached in the background (macOS/Linux only)	—
--help, -h, -?	Print CLI usage information	—
--threaded <n>	Run feature extraction using n threads	1
-nofset / -fset	Disable or force generation of the .fset file	Enabled
-nofset3 / -fset3	Disable or force generation of the .fset3 file	Enabled

Demo 🧪 (optional)

• npm run demo
• Open http://localhost:3000/
• Add -Demo to the marker command to generate demo assets automatically: node NFTMarkerCreator.js -i image.png -Demo

Python package (PyPI) 🐍

A native Python package — a pybind11 binding around the same WebARKitLib core as the Node app — is published on PyPI with prebuilt wheels, so no compiler or Docker is required.

Wheels are provided for Linux (manylinux) and Windows (AMD64), CPython 3.8–3.13.

Installation 📦

pip install nft-marker-creator

CLI 🚀

nft-marker-creator -i pinball.jpg -o ./markers --threads 4

This writes <name>.iset, <name>.fset and <name>.fset3 to the output directory. Output is byte-identical regardless of thread count.

Flag	Description	Default
-i, --input	Input image (jpg/jpeg/png)	required
-o, --output	Output directory	current dir
--dpi	Force a DPI instead of image metadata	metadata, else 150
--level	Tracking extraction level 0–4	2
--leveli	Initialization extraction level 0–3	1
--threads	Worker threads for feature extraction	1
-h, --help	Print CLI usage information	—

Python API 🐍

import nft_marker_creator as nmc

paths = nmc.create("pinball.jpg", output_dir="markers", threads=4)
print(paths)  # ['markers/pinball.iset', 'markers/pinball.fset', 'markers/pinball.fset3']

See python/README.md for the full API and details.

Additional resources 📚

• Marker quality tips — guidance on preparing images that track reliably

ES6 version of the build library

The library is built with Emscripten and is located in the build folder. It is an ES6 module and can be imported in your project.

<script type="module">

  import nftMC from '../build/NFTMarkerCreator_ES6_wasm.js';
  const mc = await nftMC();

</script>

Create your NTS markers with our docker image

First, you need docker installed in your system, if you haven't, follow the Docker engine installation instruction . Then inside the folder you want to run the docker image:

Linux/macOS:

docker run -dit --name nft-app -v "$(pwd):/src" webarkit/nft-marker-creator-app:0.4.0 bash

Windows (PowerShell):

docker run -dit --name nft-app -v "${PWD}:/src" webarkit/nft-marker-creator-app:0.4.0 bash

With the docker container generate the NFT marker:

docker exec nft-app node ../Nft-Marker-Creator-App/src/NFTMarkerCreator.js -I /src/pinball.jpg

remember to prepend the -I /src/<path to your image>

Build

Build emscripten files with docker:

Linux/macOS:

docker run --rm -v $(pwd):/src -u $(id -u):$(id -g) -e "EMSCRIPTEN=/emsdk/upstream/emscripten" emscripten/emsdk:4.0.19 npm run build-local

Windows (PowerShell):

docker run --rm -v "${PWD}:/src" -e "EMSCRIPTEN=/emsdk/upstream/emscripten" emscripten/emsdk:4.0.19 npm run build-local

or better create a docker container and run the build command inside it:

Linux/macOS:

docker run -dit --name emscripten-nft-marker-creator-app -v $(pwd):/src emscripten/emsdk:4.0.19 bash
docker exec emscripten-nft-marker-creator-app npm run build-local

Windows (PowerShell):

docker run -dit --name emscripten-nft-marker-creator-app -v "${PWD}:/src" emscripten/emsdk:4.0.19 bash
docker exec emscripten-nft-marker-creator-app npm run build-local

In VSCode you can run the setup-docker and build-docker command inside package.json.

Planned Features

• Multi threading support to speed up the creation of the markers.
• Python version of the project and app. — published on PyPI 🎉