Moodle Integration Guide: Add AI Chatbots to Your LMS

Moodle powers over 400 million users worldwide, yet most training centers still rely on manual support for learner questions. Adding an AI chatbot to Moodle can reduce support tickets by 70% while providing 24/7 assistance. This guide covers every integration method — from native plugins to LTI 1.3 — so you can choose the right approach for your organization.

1. Why Add a Chatbot to Moodle?

Moodle administrators and training managers face a common challenge: learners expect instant answers, but support teams can't scale infinitely. An AI chatbot trained on your course content bridges this gap.

The Support Problem in Moodle

What a Moodle AI Chatbot Can Do

• Answer course questions 24/7: "When is assignment 2 due?" "What topics are covered in Module 3?"
• Navigate the LMS: "Where do I find the discussion forum?" "How do I submit my project?"
• Explain course concepts: Using RAG (Retrieval-Augmented Generation), the chatbot retrieves relevant content from your materials
• Reduce admin workload: Handle password resets, enrollment questions, and policy clarifications
• Track learner engagement: Analytics on what questions learners ask most frequently

2. Integration Methods Compared

There are three main ways to add a chatbot to Moodle. Each has trade-offs between ease of setup, features, and control.

Which Method Should You Choose?

• LTI 1.3 (Recommended): The industry standard for LMS integrations. Passes user identity securely, works at course or site level, and requires no plugin installation.
• Native Plugin: Best when you need deep Moodle integration or want to self-host the AI. Requires server access and technical expertise.
• Embedded Widget: Fastest to test but limited functionality. Good for proof-of-concept before committing to full integration.

3. LTI 1.3 Integration (Step-by-Step)

LTI (Learning Tools Interoperability) 1.3 is the most secure and feature-rich way to add an external chatbot to Moodle. Here's how to set it up with Criterium (or any LTI 1.3 compatible chatbot).

Prerequisites

• Moodle 3.7+ (LTI 1.3 support required)
• Site Administrator or Manager role
• Your chatbot provider's LTI credentials (Client ID, Deployment ID, URLs)

Step 1: Configure External Tool in Moodle

1. Go to Site Administration → Plugins → Activity modules → External tool → Manage tools
2. Click "Configure a tool manually"
3. Fill in the basic settings:
   • Tool name: AI Course Assistant (or your preferred name)
   • Tool URL: Your chatbot's LTI launch URL
   • LTI version: LTI 1.3

Step 2: Configure LTI 1.3 Security

1. In the tool configuration, locate the Public keyset URL and Initiate login URL fields
2. Enter the values provided by your chatbot vendor
3. Set Default launch container to "Embed" or "New window" based on preference
4. Under Privacy, enable:
   • Share launcher's name with tool
   • Share launcher's email with tool
   • Accept grades from the tool (optional)

Step 3: Add to Courses

1. Navigate to your course and enable Edit mode
2. Click "Add an activity or resource"
3. Select "External tool"
4. Choose your configured AI chatbot from the preconfigured tools dropdown
5. Save and display to test the integration

4. Native Moodle AI Plugins

If you prefer a native plugin approach, several options exist in the Moodle plugins directory. Here's how they compare:

Plugin Installation Steps

1. Download the plugin ZIP from moodle.org/plugins
2. Go to Site Administration → Plugins → Install plugins
3. Upload the ZIP file and click "Install plugin from ZIP file"
4. Follow the configuration wizard to enter API keys and settings
5. Add the block to your dashboard or course pages

5. Training Your Chatbot on Course Content

A generic ChatGPT integration won't know your course schedule, policies, or specific content. Training the AI on your materials is what makes it actually useful.

What Content to Include

RAG: How Content Training Works

Modern AI chatbots use Retrieval-Augmented Generation (RAG) to answer questions from your content:

1. Chunking: Your documents are split into searchable segments
2. Embedding: Each chunk is converted to a vector representation
3. Retrieval: When a learner asks a question, relevant chunks are found
4. Generation: The AI uses retrieved content to formulate an accurate answer

This approach ensures the chatbot cites your actual course materials rather than making up answers — critical for educational accuracy.

6. Best Practices for Moodle Chatbots

1. Set Clear Expectations

Let learners know what the chatbot can and can't do. Include a welcome message like:

2. Monitor and Improve

• Review unanswered questions weekly and add content to fill gaps
• Track most frequent questions to identify course pain points
• Update content when syllabi or policies change

3. Enable Human Escalation

Always provide a clear path to human support. Configure the chatbot to suggest contacting the instructor or support team when it can't answer confidently.

4. Respect Privacy

• Ensure GDPR/FERPA compliance for student data
• Choose providers with EU data residency if required
• Don't train the chatbot on student submissions or grades

7. Troubleshooting Common Issues

LTI Launch Fails

• Check URLs: Ensure launch URL, keyset URL, and login URL are exactly as provided
• HTTPS required: Both Moodle and the tool must use HTTPS
• Clock sync: LTI 1.3 tokens are time-sensitive. Ensure server clocks are synchronized

Chatbot Doesn't Know Course Content

• Content not uploaded: Verify your materials were successfully processed
• Wrong course context: Ensure the chatbot is associated with the correct course
• Stale content: Re-sync if you've updated course materials recently

Users Not Identified

• Privacy settings: Check that "Share launcher's name/email" is enabled in tool config
• User role: Some tools require specific Moodle roles to pass user data

8. Frequently Asked Questions