Ask your data anything.
The Agentic Query Layer (AQL) translates natural language questions into secure, permission-aware queries against your existing systems — without exposing your data to AI models.
Core is open source. Advanced reporting tools available via modest subscription.
The first implementation is moodle-mcp-server — an open-source MCP server for Moodle LMS.
How it works
The AQL sits between your data sources and your AI models. You ask questions in plain language. The layer translates intent into safe, auditable queries, executes them against your systems, and returns structured results with context and suggested follow-ups.
Open Source Core
Install the core AQL for Moodle LMS today. Use basic functions immediately at no cost.
Advanced Reporting
Premium plugins and composite reporting tools available through modest subscriptions or self-hosted.
Enterprise Ready
Already in use by training providers, educational institutions, and government agencies.
Questions about AQL from the Workspace
Our Workspace is a great place to get your questions answered about the Agentic Query Layer — what it is and how AQL and the Moodle learning management system fit together in practice.
What is AQL? What is MCP server for Moodle?+-
AQL stands for Agentic Query Layer. AQL lets an AI agent ask questions about real system data through approved tools instead of direct database access. A Moodle MCP server is the connector that makes selected Moodle actions available to an AI client through MCP. The roles are: - Moodle stores the course, user, enrollment, grade, and activity data. - MCP lets an AI client discover and call approved tools. - Our Moodle MCP server exposes selected Moodle API actions as MCP tools. - AQL is the broader layer for controlled AI queries across Moodle and, later, other systems. For example, someone could ask: > Which learners in this course are active but falling behind? AQL routes the question through our Moodle MCP server, uses approved Moodle data, and returns structured information the AI can summarize. The AI never gets direct database access, and it never gets a chance to edit LMS data unless an approved write tool is explicitly built and enabled. Short version: AQL is the query layer. Our Moodle MCP server is the first connector.
Can Moodle connect to an AI chatbot through MCP?+-
Yes. An LMS can connect to an AI chatbot through MCP. Moodle exposes course, user, enrollment, grade, and activity data through its API. An MCP server can make selected LMS actions available to an AI client as controlled tools. The flow is simple: - The AI client asks a question. - The MCP server checks which LMS tool is allowed. - The MCP server calls the LMS API. - The LMS returns structured data. - The AI summarizes the answer. For example, a user could ask: > Which students are enrolled in Course A but have not completed the latest assignment? The AI does not need direct database access. It only gets the data exposed through approved MCP tools. This is useful for reporting, enrollment checks, course progress, support dashboards, and admin workflows.