Software
All of my software is designed to work on Linux, Mac, and Windows. If you find any of my software helpful in your research, I just ask that you please cite this webpage. Thank you!
Modular Digital Methodologies Toolkit
The Modular Digital Methodologies Toolkit (MDMT) is a comprehensive desktop application designed to streamline digital humanities research and text analysis. It brings together essential tools for researchers in one unified interface, including optical character recognition for converting scanned documents to searchable text, audio transcription, language translation, named entity recognition, relationship extraction, co-word analysis, and an AI-powered retrieval-augmented generation chatbot that draws exclusively from documents you provide. Version 2.0 is a substantial rewrite, replacing the original Tkinter GUI with a modern web-based interface built on NiceGUI and Pywebview, running in a native desktop window. MDMT's RAGBot now runs local Qwen 3.5 models and auto-detects your hardware (NVIDIA, AMD, and Apple Metal GPUs) to recommend the best model size for your system. Optional assets such as language models, Tesseract language files, and NLTK data are downloaded on demand via a built-in Downloads page rather than being bundled with the application.
If you would like to keep up with the development progress, make your own (unsupported) build, or fork this software for your own purposes, you can view the software repository here. Pull requests welcome!
Download:
Mac version 2.0.5 (Apple Silicon)
Topic Clustering Toolkit
This toolkit allows for unsupervised structuring of unstructured qualitative data by topic. The software will determine clusters of abstract topics in your corpus based on some keywords of interest. The package contains a Docker Compose file which automatically creates and configures your working environment by combining Apache Solr, Carrot2 Workbench, and some scripts I wrote to facilitate ease of use. Requires some minimal setup, but setup and usage instructions are fully documented.