Additional Services¶

Beyond text chat, uniGPT offers several additional capabilities.

Image Generation¶

Flux 1 (Current)¶

Flux 1 is currently available at https://text2image.uni-muenster.de/. It is not yet integrated into the Chat interface.

Flux 2 (Coming Soon)¶

Flux 2 will be deployed soon and integrated directly into the Chat interface, allowing you to generate images through tool calls within conversations.

Document Upload & Analysis¶

You can upload documents (such as PDFs) directly into the chat interface. The system will parse the document and allow you to ask questions about its content.

Click the 📎 attachment icon next to the chat input to upload a file.

Speech Recognition / Transcription¶

uniGPT provides automatic speech recognition (ASR) for transcribing audio and video files at https://transcribe.uni-muenster.de/. The service uses NVIDIA's Parakeet, OpenAI Whisper, and Mistral's Voxtral models, supports multiple languages, and currently accepts at least WAV, MP3, FLAC, and OGG files. The local Parakeet setup also supports additional formats, including video files. All processing runs within the University of Münster's data centers. The service is available through its own web interface and via the uniGPT API:

Translation¶

uniGPT offers a translation service that uses the official university glossary to ensure consistent terminology across all university communications.

The translation service will be available in two ways:

• Imperia integration — embedded directly into the university's CMS for seamless use when editing web content.
• Standalone service — a dedicated interface for general translation tasks, coming soon.

Integration with University Systems¶

uniGPT is integrated into several university systems:

• Content Management System (CMS / Imperia) — including upcoming translation integration
• Tutor.AI