Progress ARAG Connector brings Progress Agentic RAG to Magnolia. Built and supported by Noice, it connects your content to a Knowledge Box, adds an Ask component to your pages, and gives authors actionable AI insights.

1. Problems and challenges
Useful information is spread across pages and documents, while conventional search leaves visitors to assemble an answer themselves. Keeping a separate AI knowledge base up to date adds manual work, and authors need to understand which questions their content can answer.

2. Benefits
Visitors can ask natural-language questions and receive streaming answers with numbered links to source pages and documents. Publishing keeps the Knowledge Box current. Authors can review practical GEO and SEO suggestions in Magnolia, while the Ask component follows the site's fonts and colours.

3. How it works
Install the module on Magnolia author and public instances. In Agentic RAG settings, connect a Progress Agentic RAG Knowledge Box, choose sites and DAM folders, and load existing content. Publish, unpublish and delete events then update the Knowledge Box. Add the Ask component through the page editor. A server-side proxy keeps API credentials out of the browser.

Content sync uses published content by default; administrators can change this setting. AI insights analyses the page selected by an author and sends that page's text to the configured Knowledge Box, including unpublished edits when selected. Insights are suggestions and do not change pages automatically.

4. Key features

• Publishing sync for pages and DAM documents, plus resumable bulk loading with retries and visible failures.
• Responsive Ask component with streaming answers, numbered citations and suggested questions.
• Author-controlled GEO/SEO analysis, citation self-tests and site-level reports.
• Native Magnolia authoring, including an Analyse with AI insights action in the Pages app.
• Server-side credentials, encrypted or environment-based key storage, and dedicated administrator/editor roles.

Pricing and support
The module is free under Apache-2.0. Implementation and support from Noice are paid services, quoted separately. Progress Agentic RAG licensing and usage charges are separate.

Compatibility and prerequisites
Module version 1.0.0. Tested with Magnolia 6.4.10 Community Edition and DX Core, Java 17 and Tomcat 10.1. DX Cloud has not yet been tested. Magnolia 6.2 is unsupported.
Use a vendor-supported, security-patched Magnolia 6.4 webapp and supported Java 17 runtime.
A Progress Agentic RAG Knowledge Box and appropriate API keys are required. Progress Agentic RAG licensing and usage charges are separate.

Module access, implementation and support: jake.tracey@noice.net.au.