Decthings Documentation

Learn and discover everything you need to get the most out of the platform.

API clients and reference

Call Decthings functions from your own code. Make evaluations to allow your application to make intelligent decisions.

What is Decthings?

Artificial intelligence (AI) allows a computer program to make intelligent decisions based on data. In the end, AI is useful to us humans because we can connect it to automated systems. For example, an AI could monitor a construction site and alert us humans if it sees a construction worker in danger. It could also be used for our daily tasks, such as for automatically switching on the lights when we enter a room, or for generating text-to-speech for automated audio books.

Decthings is a platform for running artificial intelligence (AI) in the cloud, along with tools to create and train your AI. There are several advantages to this approach:

Creating a model is easier with tools built for AI. You can train your model with the click of a button, your input data will automatically be loaded, and debugging allows you test that everything works as expected.
You don't need your own hardware. You can train on powerful cloud hardware, and you can remotely evaluate your model from any device with a network connection.
Deploying a finished model to your devices is simplified. You can manage your AI from the platform instead of manually configuring each device.

In Decthings, a model is a component that can take some input data and transform it into some output data. This can be both a simple algorithm or a complex neural network. Below is an example of an image segmentation model, which outputs a new image where different objects are marked in color.

The evaluation user interface, where an image has been provided to the Deeplab image segmentation model.

Remotely evaluate your model from your device

Use an API client to remotely make an evaluation with a few lines of code. We aim to create API clients for more languages in the future.

Language

import { DecthingsClient, Data, DataElement } from '@decthings/api-client';

async function main() {
    let apiKey = fs.readFileSync('./auth.txt')
    let client = new DecthingsClient({ apiKey });

    // Construct the input data, which in this example is just an input text
    let inputData = new Data([DataElement.string("My input text")])

    // Evaluate the model. Replace with your model ID
    let result = await client.model.evaluate('<model_id>', [{ name: 'input', data: inputData }]);

    // Do something with the result!
    console.log(result);
}
main()

Starting point for beginners in AI and programming

Glad to have you on board! We hope that by providing you with useful tools we can make it easier to learn about AI and programming in general. You can get started easily and learn as you go.

Without writing any code, you can build an entire AI-based automated system. You can for example control internet-of-things devices such as smart lights and other home appliances. This can be done by using the visual editor to create a model, and the Node RED system to create your application. Node RED is a tool for building automated systems without writing code, and with Decthings you can use AI from within that system. Pretty cool! Take a look at the Visual editor and Node RED guide to learn more.

Starting point for experienced developers

By providing features that are only possible on a cloud-based platform we hope that experienced developers can find the platform useful as well. One such feature is to make it much easier to connect and manage many devices. Instead of manually deploying your code to your devices you can with one click update them to use the latest and most optimized model.

Changing workflow, IDE or other tools is a pretty difficult thing to get a developer to do. So don't! We want you to be able to seamlessly transition to Decthings. For example, we support several languages to create a model, with more coming soon. You can also use your preferred editor, such as Visual studio Code or any other IDE. You can even upload your existing model to Decthings (with some adjustments). We try to make the platform as flexible as possible in order to support as many different use cases as possible.

To get started, take a look at the Visual editor guide and Code an image classifier guide.

These are early stages for Decthings

Decthings is in development. We believe the core functionality is stable, but things may still change at any time.

We are happy to have you along, and we are striving towards producing a stable API and platform that can be trusted for ciritcal applications, but for now, use Decthings for non-critical applications only.

This also means that we are really thankful for any feedback and feature requests you can provide. We are excited about new features and about making the platform as enjoyable as possible. If you have questions, want to request a feature or report a bug or issue, contact support@decthings.com.

We hope that you have a hassle free experience using Decthings, and hope that you stay with us as we make the platform better.