eLeDia.ai RagIngest

eLeDia.ai RagIngest prepares Moodle course content for retrieval-augmented generation, or RAG. The plugin extracts selected Moodle course materials, converts them into structured documents and sends them to a configured RAG service, where they can be indexed for AI tutors, Moodle chatbots and LLM-based learning assistants.

The result is a Moodle AI experience that is grounded in real course content. Instead of relying only on general model knowledge, an AI tutor can answer questions using the actual materials, activities and resources from released Moodle courses. This helps learners receive more relevant answers and gives institutions a controlled way to make Moodle content available for AI-supported learning.

Moodle RAG Indexing for Course Content

RAG depends on reliable source material. eLeDia.ai RagIngest is designed to collect that material from Moodle in a controlled and repeatable way. The plugin extracts content from supported Moodle activities and resources, prepares it for ingestion and sends it to an external RAG endpoint.

This allows a connected AI tutor or LLM system to retrieve relevant course passages when answering questions. Answers can then be grounded in Moodle course materials instead of being generated only from the model’s general training data.

For education, this is essential. Learners do not just need plausible answers. They need answers aligned with their actual course, their teacher’s materials and the institution’s learning context. eLeDia.ai RagIngest helps provide that foundation.

Controlled Opt-In Indexing for Moodle Courses

eLeDia.ai RagIngest is built around opt-in indexing. No course is sent to the RAG service unless it has been explicitly released.

Administrators can release courses through a central pilot-course list, a category allow-list or a course-level override. This makes it possible to start with a controlled pilot, expand gradually to selected categories and still handle individual exceptions when needed.

During test phases, course marking can be locked so teachers cannot accidentally change which courses are indexed. Existing course-level values are preserved and can become active again when the lock is removed.

This opt-in approach gives institutions the control they need when introducing AI and RAG into Moodle. Administrators decide which content becomes part of the AI knowledge base and when.

Activity-Specific Moodle Content Extractors

Moodle courses contain many different types of content. A page, quiz, book, folder, glossary, H5P activity and assignment all store learning material differently. eLeDia.ai RagIngest handles this through activity-specific extractors.

The plugin ships with extractors for common Moodle activity and resource types, including assignments, books, databases, feedback activities, folders, glossaries, H5P activities, IMS content packages, labels, lessons, pages, quizzes, files, SCORM packages, Video Time activities, wikis and workshops.

Each extractor focuses on the educational content that should be available for retrieval. For example, assignment instructions and grading criteria can be indexed, while learner submissions are not included. Feedback activity questions can be indexed without sending submitted responses. Folder modules can create multiple documents from supported files. Video captions can become searchable learning content.

This structured extraction makes Moodle content more useful for AI than a simple page scrape. The RAG service receives meaningful documents with source information, content type and Moodle metadata.

Grounded AI Answers for Moodle Learners

When eLeDia.ai RagIngest is used with eLeDia.ai Tutor or another RAG-enabled assistant, learners can ask questions that are answered against actual course materials.

A learner might ask for an explanation of a topic, help finding a relevant resource, a summary of a course section or guidance on a quiz concept. The AI tutor can retrieve matching Moodle content from the RAG index and use it to generate a response.

This reduces hallucination risk and makes the assistant more useful in real teaching and learning contexts. The AI is not just a generic chatbot. It becomes a course-aware learning assistant that can refer to the institution’s own Moodle content.

AI Support for Teachers and Course Teams

RAG indexing is also valuable for teachers, course designers and support teams. Once course content is indexed, AI tools can help search, summarise and explain materials more effectively.

Teachers can benefit from an assistant that understands the structure and content of their courses. Course teams can pilot AI support in selected courses before scaling across departments. Administrators can see which released courses are waiting for indexing and queue indexing tasks directly from the plugin interface.

Because indexing is controlled centrally, institutions can introduce AI features without losing oversight of which materials are included.

Automatic Moodle Content Ingestion

eLeDia.ai RagIngest supports automatic ingestion through Moodle events. When supported course modules are created, updated or deleted, the plugin can queue ad-hoc tasks to update the RAG index.

This helps keep the AI knowledge base aligned with Moodle. If a teacher updates a page, changes a book chapter, modifies a quiz, adds glossary entries or changes supported activity content, the plugin can queue the relevant indexing work.

Deletion is handled as well. When a module is removed, the corresponding source can be deleted from the RAG index, helping prevent outdated content from remaining available to AI systems.

Privacy-Aware Moodle AI Indexing

eLeDia.ai RagIngest is designed to index course content, not user-scoped personal learning records. The plugin declares that course content and module metadata are transferred to the configured RAG service for parsing, chunking and indexing.

Extractors are careful about what they include. For example, learner submissions and feedback responses are excluded where they should not become part of the shared AI knowledge base. Oversized files can be skipped or content can be truncated according to configured limits.

This supports a more privacy-aware approach to Moodle AI integration. Institutions can make learning materials available for RAG while keeping control over the scope of indexed content.