Welcome to Text Lab
Text Lab is an interactive platform designed to make advanced Natural Language Processing (NLP) tools accessible to researchers and students at the University of Bern.
Developed by the Data Science Lab (DSL), Text Lab allows you to process documents, audio, and data without needing to write code.
Data Privacy & Security
Why use Text Lab?
One of the primary advantages of Text Lab is data sovereignty
- Internal Processing: All processing happens within the University of Bern's infrastructure: on the UBELIX compute node of your session, or, only if you choose the GPUStack option in the Knowledge Graph, on the university's GPUStack service.
- No Third Parties: Unlike commercial cloud services (like ChatGPT or Google Cloud), your data is never sent to external servers.
- Nothing Left Behind: Files you upload and intermediate results are kept in a private workspace for your session, which only your user account can open, and are deleted when the session ends. What you keep is up to you: the files you download, and the output folder you choose in the Knowledge Graph. Each session also keeps technical log files in your home directory (see Reporting a Problem).
- Sensitive Data: Text Lab is the better choice for working with potentially sensitive research data (e.g., for interviews and/or non-public documents).
- No coding required Text lab offers advanced NLP models and tools without having to write code.
Available Tools
- Transcription: Convert audio files to text with high accuracy, featuring a specialized model for Swiss German.
- OCR (Optical Character Recognition): Extract text from scanned PDFs using state-of-the-art AI models.
- Translate: Securely translate text and complex documents while preserving formatting.
- Meeting Notes Generator: Automatically transcribe and summarize audio recordings into structured notes, with each point linked to the part of the transcript it comes from.
- Chat: Interact with local Large Language Models (LLMs) securely.
- Data Visualization: Generate insights and plots from your data.
- Knowledge Graphs: Extract metadata from research papers and visulise the data with knowledge graphs.
- Topic Modeling: Apply unsupervised algorithem to raw text data to discover topics and trends.
For processing large amount of data, please contact the Data Science Lab.
Getting Support
If you encounter issues, need any data science, research IT support, or have questions about using Text Lab, please contact the Data Science Lab. When reporting a problem, please attach the session's log files (how to find them): support.dsl@unibe.ch
If you'd like to be informed about updates and changes to Text Lab, please subscribe to the following Mailing List
General Disclaimer (Applies to All Features)
The use of Text Lab and its featured tools is provided "AS IS."
While we strive for accuracy, most of the results are AI-generated and are intended for informational and drafting purposes only.
Users are solely responsible for verifying all data, transcripts, translations, and summaries before relying on them for critical decisions, legal action, or professional reporting.