UniGPT: New Version with Expanded Functions and Improved Interface
For almost two years now, employees and students at the university have been able to use UniGPT, an AI service developed specifically for the University of Münster. A new version has recently been released, offering users new functions and a fresh new look. "The new version offers many advantages, for example in terms of digital accessibility, because the new interface has a read-aloud function," explains Jonathan Radas, who is primarily responsible for the conceptual development of UniGPT at the CIT. In addition, it is now possible to edit prompts retrospectively, i.e., after they have been sent. Another new feature is the speech-to-text function, which allows queries to be made via voice input.
Moreover, the new version of UniGPT offers additional functions for more individualized and efficient use: for example, users can save their own prompts and thus create a personal prompt catalog for recurring tasks. Advanced settings allow for finer adjustment of responses, for example in terms of creativity or style. Uploaded files are now managed and remain available for later sessions. Another new feature is the ability to use two models in parallel to compare results directly. Current models such as Google Gemini 3 and Anthropic Claude 4 are supported, among others.
Overall, the new version is designed to make UniGPT clearer and easier to use. "Personal settings can now be conveniently customized in detail according to individual preferences," emphasizes the IT expert. For example, users can choose the language, font size, and design (light/dark) as well as an AI voice for the read-aloud function. Elements can also be shown or hidden. To familiarize yourself with the changes and new features, there is an e-learning course in Confluence.
If you have any questions, we are happy to help you via the UniGPT Mattermost channel or by phone via our hotline.