User interfaces AI assistant for cross-platform UI development

Related Vendors

The Qt Company

FAULHABER Drive Systems

Taiwan Machine Tool & Accessory Builders' Association (TMBA)

Qt users can now integrate large language models (LLMs) of their choice, including self-hosted models, into their workflow to reduce the time spent on everyday tasks in cross-platform development.

The Qt Group has announced the availability of the Qt AI Assistant. The AI assistant serves as an experimental tool to streamline the cross-platform development of user interfaces (UI). Its goal is to reduce repetitive tasks in UI development that often do not require programming. This allows developers to spend more time on the productive and creative aspects of writing innovative code. The tool supports self-hosted language models (via the cloud or locally), without the need for a third-party provider for AI support.

The Qt AI Assistant is designed to advise developers on creating application interfaces with the Qt Framework. Additionally, the tool automates manual tasks such as writing unit test cases, code documentation, and repetitive code. Users request this guidance through a command prompt window and quick access commands in the Qt Creator code editor (which is used for creating and modifying cross-platform applications).

The AI-powered assistant is compatible with various programming languages such as C++ and Python, but it has been specifically trained on thousands of real-world use cases of QML and Qt Quick. These are the language and toolkit of the Qt Framework for designing the appearance and behavior of applications. This means developers can, for instance, generate the necessary repetitive QML code while creating user interfaces, thereby gaining time for more complex programming tasks.

The Qt AI Assistant supports any language model that developers wish to use. The newly released first version is preconfigured to immediately support the following models:

• Claude 3.5 Sonnet (chat + completions)

• OpenAI GPT-4o (chat + completions)

• Llama 3.3 70B QML (chat)

• Code Llama 13B QML (code completions)

• Code Llama 7B (code completions)

"The industry has been overwhelmed with AI announcements—everyone has their preferred language model, and we don't want to take that away from anyone," says Peter Schneider, Senior Product Lead at the Qt Group. "It was important that we support an open approach and give companies the freedom to choose their preferred LLM deployment method."

The flexibility of the Qt AI Assistant allows different content creation requirements to be routed separately to different models. For instance, a developer can forward QML code to a specialized LLM while directing other code to a different LLM. In this way, developers always receive code suggestions from the most powerful models for their programming language and not from the LLM available through a third-party AI assistant provider.

The tool also promises to address security concerns for embedded devices and high-end desktop applications developed with the Qt Framework. By supporting self-hosted models, companies could carry out private cloud deployments of LLMs. This allows for the prevention of code leaks and better protection of intellectual property. Developers can even choose to trigger the Qt AI Assistant manually to avoid unwanted interruptions during programming.

Support for additional models is planned for the near future. Later this year, the Qt Group intends to release more fine-tuned versions of LLMs—starting with Llama 3.3 70B and Code Llama 13B—which have been trained on more than 4,000 human-created and validated QML code snippets. These will be available for free download on the Hugging Face portal, even for open-source Qt developers, to connect them with third-party assistants.