Midrange Dynamics Midrange Dynamics

Enterprise AI Platform for IBM i

Knowledge Driven Change

Midrange Dynamics enables organisations to safely use AI to understand, change and modernise IBM i applications — without exposing entire source code repositories or allowing AI to make uncontrolled changes.

The platform combines MDiQ, secure REST APIs, an MCP Extension and MDChange. Together these provide an AI platform that not only answers questions but can safely participate in governed software delivery.

Request Demo See the architecture

The Simple Architecture

AI Assistant / Custom Client

ChatGPT, Claude, Copilot, Cursor, Windsurf, other agents, or any REST consumer

MCP Server / Extension

Agent requests from the IDE or standalone AI agents

Secure REST APIs

Knowledge + DevOps governance — also callable directly by custom clients

MDiQ

Application Knowledge Graph

MDChange

Change Governance & DevOps

IBM i Applications

Unlike traditional analysis tools

The Midrange Dynamics platform combines structured knowledge, controlled APIs, IDE-native AI access and enterprise DevOps — so AI never needs unrestricted access to source code or production systems.

MDiQ

A structured application knowledge graph, MCP server and application explorer.

Secure REST APIs

Controlled access to application knowledge and DevOps governance, used by the MCP server for agent requests and available for direct access from any custom client.

MCP Extension

Seamless AI integration with IDEs such as VS Code, Cursor, Windsurf and IBM's BOB tooling.

MDChange

Enterprise DevOps governance that controls every change from development through deployment.

Every request — whether it comes from an MCP agent or a custom client — flows through secure, controlled REST APIs that expose only the required knowledge and governance capabilities.

Why this matters

Most AI tools work by exposing large amounts of source code and asking an LLM to infer relationships. MDiQ takes a completely different approach.

Before AI is involved, MDiQ analyses the application and builds a node-edge knowledge graph describing the application. AI therefore receives structured business knowledge rather than having to discover it every time.

This significantly improves:

  • Accuracy
  • Consistency
  • Speed
  • Explainability
  • Token efficiency
  • Scale on very large applications

Knowledge graph entities

Library PhysicalFile LogicalFile Index Column Trigger Constraint SqlFunction SqlProcedure Domain DomainRule RecordFormat DisplayField ILEModule ILEProcedure ServiceProgram Program Command DisplayFile ProgramParameter ProgramVariable Subroutine BusinessRule DataModelRelationship

Secure knowledge and governance exposure

The knowledge graph and DevOps governance layer are never exposed directly. Instead they are accessed through carefully designed REST APIs.

The MCP server

Maps selected APIs to MCP tools so AI agents can request only the information or governance action needed for a question or change.

Any custom client

Can call the same APIs directly — portals, internal tools, automation scripts, or other enterprise systems — without going through an IDE or an LLM.

APIs expose only the metadata or governance capability required

The same REST surface can also surface MDChange governance — change status, impact of a proposed deployment, required tests, approvals and audit context — so both knowledge and proven DevOps control are available through one controlled interface.

  • Which programs update Customer Balance?
  • Show every object affected by this field.
  • Explain this RPG procedure.
  • Which APIs call this program?
  • What must be tested before deployment?
  • What fields are used in the program?
  • What variables does this database field relate to?
  • Where is this procedure explicitly used?
  • What variables are related to a specific variable, either directly or indirectly through cascading assignments across programs?

This dramatically reduces unnecessary AI context while improving answer quality, and gives non-AI clients the same structured access without a separate integration stack.

AI integration through MCP

The same knowledge and governance APIs are available inside modern AI-enabled IDEs through an MCP Extension. The MCP server issues agent requests against those REST APIs; it does not bypass them.

Developers simply ask natural language questions. The MCP layer securely retrieves only the required information via the REST APIs and returns structured answers to the AI.

MDiQ MCP Server will be released in Q4 2026 with a centralised server version along with an on-prem LLM model architecture, avoiding any need for cloud-based AI LLMs or internet interaction.

This means that both IDE and other non-IDE AI agents can access the Application Knowledge and DevOps Governance provided by MDiQ and MDChange respectively — through the same REST APIs that custom clients can call directly. This will add significant security, cost and performance benefits for large scale development or modernisation projects that use AI Code Generation.

Supported environments

  • VS Code
  • Cursor
  • Windsurf
  • IBM BOB environments
  • GitHub Copilot
  • Claude Code
  • Future MCP-compatible IDEs

Why MDChange makes the difference

Many AI tools stop after providing recommendations. MDChange goes much further.

Because the AI platform is integrated with MDChange, approved AI-generated work can participate in governed DevOps processes. Those governance capabilities are exposed through the same REST APIs used for knowledge queries, so MCP agents and custom clients operate against proven MDChange controls rather than a parallel, ungoverned path.

Checkout and locking

Build orchestration

Dependency-aware compilation

Testing

Approvals

Rollback

Deployment

Audit trails

Release management

AI therefore operates within established governance rather than bypassing it.

The result is higher quality, repeatability and significantly lower operational risk.

Business benefits

Better AI answers

Structured application knowledge eliminates guesswork.

Faster modernisation

Developers spend less time discovering how applications work.

Reduced risk

Only approved, governed processes reach production.

Higher productivity

AI becomes an active engineering assistant instead of merely a chatbot.

Enterprise scale

Large IBM i applications can be analysed without sending entire codebases to an LLM.

Open but controlled access

The same REST APIs that serve MCP agents can be called directly by custom clients, so knowledge and DevOps governance are reusable across the enterprise without exposing source or loosening control.

Comparison with traditional analysis tools

Traditional products are excellent for documentation, visual exploration and impact analysis. The Midrange Dynamics approach extends beyond documentation by making application knowledge and DevOps governance consumable through secure REST APIs, integrating directly into developer workflows via MCP, and governing change through enterprise DevOps.

Capability Midrange Dynamics X-Analysis / Arcad Observer / Ozcar
Interactive application knowledge graph Yes Primarily documentation focused
REST APIs exposing structured knowledge and DevOps governance Yes — MCP agents and custom clients Limited or product specific
MCP integration for AI agents Yes Generally not a core capability
AI-ready structured knowledge Yes Primarily visual navigation
DevOps governance integration Yes — MDChange Typically external or limited
Controlled AI-assisted deployments Yes Not a primary objective

Competitive position

Rather than replacing documentation tools, the Midrange Dynamics platform creates an enterprise AI layer above IBM i applications.

  • Structured knowledge
  • Secure REST API access for MCP agents and custom clients
  • Native stored procedure knowledge access
  • AI integration
  • Governed software delivery

This enables organisations to move from understanding software to allowing AI to safely participate in software engineering.

Key sales message

MDiQ gives AI the knowledge. Secure REST APIs expose that knowledge and MDChange DevOps governance to the MCP server (for agent requests) and to any custom client (for direct access). The MCP Extension and Server bring it into the developer's IDE or AI agents. MDChange ensures every AI-assisted change is governed, repeatable and production-ready.

MDiQ — key product components

  • Builds a complete, subject matter expert map of all application entities on the IBM i

  • An application repository explorer UI implemented for both IDE Extensions (Bob, VS Code, Cursor and others) and a pure web UI

  • Enables creation and management of application domains and subdomains for scoping projects or knowledge content

  • Markdown documentation server for building and sharing business and technical documentation around the extracted knowledge graphs — with PDF, Word and Spreadsheet exports

  • MCP Local IDE extension and IBM i centric MCP Server (Q4 2026)

  • Secure REST APIs that expose MDiQ knowledge and MDChange DevOps governance to the MCP server and to any custom client

  • SQL Stored Procedures for exposing knowledge graphs natively on IBM i to any IBM i process or program

MDiQ development roadmap

Integration with MDTest

Exploiting AI CodeGen to auto-create MDTest unit tests in RPGLE.

Knowledge Driven Change in MDChange

Real-time end to end visibility of change projects.

Sitemule modernisation integration

Re-engineering applications into modern UIs with REST API backend generation, using AI governed by MDiQ knowledge graphs.

Speak To Our Experts

For more info, a live demonstration, or to discuss governed AI for your IBM i applications with the Midrange Dynamics team.

Request Info