Build AI-Powered Experiences with Mendix

In a single platform, Mendix offers everything you need to build AI-augmented applications.

Try Mendix for free

 

trusted by

Create context-aware, proactive apps
in a low-code way

Meet Maia for Smart Apps

With a full spectrum of AI capabilities, Maia for Smart Apps makes it easy to deliver the most engaging user experiences for your intelligent applications.

Our platform abstracts the complexities of building AI-powered applications, and is the only one of its kind with the versatility to support external generative AI and cognitive services, as well as embedded ML models.


Make every application a smart application

With our generative AI capabilities, you can use the model of your choice and easily implement your own data in a controlled manner.

Generate new content

Make internal knowledge bases searchable to find information faster and create new content such as summaries, reports, articles, etc.

Get data-driven insights

Leverage enhanced decision-making and analyze large dataset to discover trends or anomalies obvious to humans.

Improve core processes

Generative AI improves processes by automating repetitive tasks, enhancing decision-making, and optimizing workflows.

Create conversational experiences

Mendix makes it easy to add conversational UI to your applications using any supported Large Language Model (LLM). You can also integrate a knowledge base to ground the responses of the LLM using your own data.

Fully integrated with our Bedrock and OpenAI connectors, our Conversational UI provides everything you need to get started and quickly deliver an intelligent chat experience.

  • Leverage your own data to provide grounded answers using knowledge bases like those supported by Amazon Bedrock, or create your own with the PgVector extension for PostgreSQL.

  • Our best-of-breed UI/UX enables customers to deliver AI-generated information in various ways, from web and native mobile applications to AR/VR and chat assistants.

Connect to any AI service

Use our built-in connectors for industry-leading AI services like OpenAI‘s GTP-3.5, GPT-4, and Amazon Bedrock to deliver intelligent solutions through an integrated developer existence.

Powered by best-in-breed AI models and services and a strong partnership with AWS, Mendix Connectors help you create apps that can:

  • Analyze and comprehend large amounts of text
  • Extract information from documents, images, and videos
  • Localize context by translating text into multiple languages
  • Synthesize text into speech for speech-activated experiences

In the rare case that we don’t have a connector, use our normal REST APIs to interface with any AI service.

See all connectors in Marketplace

Embed ML models directly into applications

Build models using any ML framework and easily integrate them into your Mendix applications. Mendix is the first and only low-code platform to support the embedded integration of ML models directly into your application, rather than using third-party hosted services.

Our Machine Learning Kit lets you add a trained ONNX-compatible model into your application and interact with it like a first-class citizen.

The AI model runs locally inside the Mendix runtime environment, which provides three key benefits:

  • Lower cost

    Avoid services that charge usage-based pricing.

  • Better performance

    Reduce the latency of calling out to an external host to process your request.

  • Increased security

    Keep sensitive data in your runtime environment to avoid sending it to an external host.

Related resources

FAQs

What benefits do embedded ML models provide over hosted AI services?

When you embed and AI/ML model inside your Mendix application, your model and application run inside the same node. This provides many benefits when compared to hosted third-party AI services, such as:

  • Better performance
  • Less latency
  • Lower transactional cost
  • Better data security and privacy
  • Faster time to market
Where can I find pre-trained models for my smart applications?

You can see what is available in the ONNX Zoo, which is a collection of pre-trained, state-of-the-art models in the ONNX format contributed by its community members.

If you can’t find what you are looking for in the ONNX Zoo, then any model created by TensorFlow, PyTorch, Caffe, Keras, Wolfram, spaCy, or other AI framework that can convert their models into ONNX format can be used withing your Mendix applications.

Many models are also available on Hugging Face.

Am I limited to using only one type of AI capabilities per application (embedded or hosted)?

No, you can use a combination of embedded and hosted AI capabilities.

Mendix provides a unified, integrated developer experience with a wealth of AI services and models at your fingertips. Compose solutions from a multitude of AI capabilities to create new and innovative digital experiences for customers and employees.