Southwark Council and Chicken Develop AI-Powered PDF Importer for LocalGov Drupal
Southwark Council and digital agency Chicken are developing an AI-powered PDF importer for the LocalGov Publications Module. The goal: convert static documents into structured, accessible HTML in under a minute—dramatically accelerating content workflows for councils.
Using a modular ETL pipeline, the importer extracts text from PDFs, transforms it using the Claude Sonnet AI model, and saves it as HTML-ready Drupal content. It supports custom import pipelines and a plugin-based architecture, allowing councils to adapt the tool for various document types, AI prompts, and content types. Early results show successful handling of tables, images, heading hierarchies, and pagination based on logical content flow.
This initiative—part of the Drupal AI and LocalGov efforts—exemplifies how public sector teams can use open source and generative AI to reduce publishing friction and improve accessibility at scale. Once released, the importer will be available for reuse across the LocalGov Drupal network. The project was outlined in a guest blog post by Angie Forson, Web and Digital Programme Lead at Southwark Council, published on the Drupal AI Initiative blog.
