IWAC AI Pipelines

Python pipelines that apply large language models to the curation of a digital archive: OCR extraction and correction, summarization, named-entity recognition with authority reconciliation, audio and video transcription, handwritten text recognition, and sentiment analysis, all written back to an Omeka S instance. Built for the Islam West Africa Collection (IWAC) and adaptable to other digital humanities and social science collections. If you use these AI workflows in your research or projects, please cite the repository as below.