Without structured knowledge
AI rediscovers, every time
- Question
- AI Agent
- Find source
- Read source
- Infer relationships
- Find more source
- Infer dependencies
- Reason
- Answer
- Repeated discovery
- Large context
- Higher cost
- Incomplete knowledge
Why MDiQ?
Don’t make AI rediscover your application.
Without structured application knowledge, an AI agent may need to rediscover how your application fits together from source code every time you ask a question.
With MDiQ
AI starts from known application knowledge
- Question
- AI Agent
- MDiQ
- Relevant Knowledge Graph
- Reason
- Answer
MDiQ supplies the structure already parsed, mapped and connected — and automatically kept current through CI/CD triggers.
- Structured context
- Faster answers
- Lower AI workload
- Known relationships
AI integration
Give your AI agent the application knowledge it needs.
Use the AI agent you choose.
MDiQ provides AI agents with structured, deterministic knowledge of IBM i applications through Model Context Protocol (MCP).
Instead of requiring an agent to discover application structure from source code, MDiQ can provide the relevant programs, procedures, files, fields, variables, parameters, dependencies and relationships as targeted structured knowledge.
- IBM Bob
- Claude
- GitHub Copilot
- Cursor
- Other MCP-compatible AI agents
Designed to integrate with MCP-capable AI agents and development environments, including IBM Bob and other compatible tools.
Independent of the AI agent
MDiQ provides the application intelligence layer independently of the AI agent consuming it. MCP is a major integration mechanism, but REST APIs and SQL Stored Procedures remain available — so MDiQ is never tied to a single AI vendor or model.
Bring AI to your application — then give it precisely the knowledge it needs.
Even when AI is deployed locally, MDiQ reduces the need to repeatedly load, search and interpret large source repositories.
- Smaller AI context
- Less repeated discovery
- Faster analysis
- Lower inference workload
- Deterministic relationships
- Reusable application intelligence
- Controlled exposure of relevant knowledge
The technology that powers MDiQ
What is an Application Knowledge Graph?
MDiQ builds the cross-reference you always wished you had: a comprehensive node-labelled, directed relationship map of your IBM i application that captures not just source code, but the complete application structure and the relationships between its assets. It is not a list of objects or a source-code search index.
Programs, service programs, modules, procedures, database tables, files, fields, variables, parameters, subroutines, displays and other application entities are represented as interconnected nodes — creating a precise structural map of what exists, how it is related and how the application works as a whole.
This gives developers — and AI agents — a reliable way to navigate the application from any starting point, follow dependencies and understand impact across the system.
Instead of asking AI to guess how a legacy application works from isolated pieces of source code, MDiQ gives it the map of the application first.
Built on IBM i expertise
Go deeper than object dependencies.
Understanding which IBM i objects depend on each other is only the beginning.
MDiQ’s native parsers analyse application source and object metadata to understand relationships inside those objects — and the knowledge graph is automatically kept current through CI/CD triggers.
The parsers were developed by experienced IBM i experts — including contributors to IBM Redbooks and authors of commercial legacy analysis products.
- Procedures inside programs and service programs
- Variables
- Database fields
- Calling and called parameters
- File usage
- Field usage
- Cross-program parameter flow
- Data flow
- Subroutines
- Display usage
- Multi-level dependencies
Native Parsing
Accurate and comprehensive
MDiQ uses IBM i-native parsing to extract and understand your application’s real structure — not just text, but meaning.
What MDiQ captures
- Database fields
- Program variables
- Procedure variables
- Constants
- Device references
- Calling and called parameters
- Source members
- Object metadata
- File definitions
- Display files
- Dependencies
- And more...
Intelligent Variable Mapping
Real application context
Parsers analyse source code and object metadata to map variable relationships across IBM i applications.
Not just where variables exist
- Operational context
- How they interact with other variables and fields
- Data flow relationships
- Cross-program parameter mapping
- Multi-level dependency chains
Program
Procedure
Service Program
Module
Advanced ILE Mapping
True execution insight
Visibility into the actual logical execution elements inside service programs and modular applications — not only the object that contains them.
Examples of relationships captured
Program → Program
Program → Procedure
Program → Module
Module → Procedure
Service Program → Procedure
Procedure → Module
Procedure → Service Program
Service Program → Module
Call structures are captured between programs, procedures and modules in every direction — program-to-program, program-to-procedure, procedure-to-program and procedure-to-procedure.
One Knowledge Graph. Every way you need it.
Application knowledge for people, tools and AI.
MDiQ turns your IBM i application into a reusable knowledge service. Developers can explore it visually. AI agents can query it through MCP. Enterprise applications and development tools can consume it through REST APIs or native IBM i SQL Stored Procedures.
MDiQ · Application Intelligence
Application Knowledge Graph
-
Graphical Interface
For people
Explore the application through purpose-built views, dashboards, call charts and where-used analysis.
-
MCP
For AI agents
Give MCP-capable AI agents targeted, structured application knowledge instead of raw source.
-
REST APIs
For tools
Governed access to the Knowledge Graph as structured JSON for custom clients and third-party tools.
-
SQL Stored Procedures
For IBM i applications
Query the Knowledge Graph natively on IBM i from SQL, your own programs and existing tooling.
One authoritative source of application knowledge — available wherever you need it.
One Knowledge Graph. Many uses.
Analyse • Visualise • Document • Govern — turn application knowledge into insight, documentation and governed change.
Analyse
Metrics, impact analysis and change scoping.
Get detailed metrics on your application — including object counts, complexity, dependencies and code distribution — to scope impact, estimate effort and identify modernization priorities.
Repository Metrics
Visualise
Interactive application maps and object dashboards.
Explore interactive diagrams that show how programs, files, procedures and other objects are connected. Follow call relationships and data usage, trace dependencies and see the bigger picture.
Object Dashboard
Document
Generate and maintain technical documentation from authoritative application knowledge.
Export structured documentation directly from your Application Knowledge Graph to a web-based, Markdown-based platform and edit it in a modern web interface. Keep your documentation current, searchable and linked to real application objects.
MDiQ Docs
Govern
Extend application understanding into controlled software change through MDChange integration.
Go from insight to action. Use the Knowledge Graph to check out related objects, manage changes, compile, implement and deploy — all within proven DevOps governance.
MDChange
Governed AI
From AI understanding to governed change.
Application understanding is only the first part of safely using AI for enterprise software development. MDiQ supplies authoritative application knowledge, while MDChange keeps AI-assisted change within your established DevOps controls.
AI reasons.
MDiQ knows.
MDChange governs.
-
AI agent
Reasons about the task, using the AI tool you choose.
-
MDiQ — Application Intelligence
Provides the authoritative knowledge needed to understand impact.
-
MDChange — DevOps Governance
Controls what is permitted to change, and how it reaches production.
Governed by MDChange
- Checkout and controlled source access
- Build
- Test
- Approval
- Deploy
- Audit
- Rollback
Governance and controlled actions are central: an AI agent does not receive unrestricted authority to modify or deploy software — its work participates in the same DevOps controls as any other change.
KEY BENEFITS
|
KEY BENEFITS
|
KEY BENEFITS
Native on IBM i
Complete node-edge mapping with native parsers on the IBM i server — not an off-platform reconstruction of the application.
Always current
CI/CD triggers granular repository updates so the Knowledge Graph stays aligned with the live application.
Procedure-level ILE
Trace calls and file usage to the exact ILE procedure — not only the service program that contains it.
Accurate impact analysis
Intelligent variable mapping, multi-level dependency chains and where-used tracing across even the largest IBM i systems.
Grounded application knowledge
Known application relationships instead of AI inference — reducing structural hallucination by giving AI authoritative relationships rather than asking it to infer them repeatedly from source.
AI-ready application intelligence
Give AI the knowledge it needs — not the entire source tree. Targeted, structured context through MCP, REST and SQL means smaller context, less repeated discovery and lower inference workload.
Use the AI agent you choose
MDiQ provides the application intelligence layer independently of the AI agent consuming it — designed for IBM Bob and other MCP-capable agents, with REST and SQL always available.
Bring AI to your application
Even when AI runs in your own environment, MDiQ reduces the need to load, search and interpret large source repositories — and controls which relevant knowledge is exposed.
Object-linked docs
MDDocs associates Markdown documentation with programs, procedures and database elements — and exports to PDF and Word.
Speak To Our Experts
For more info, a live demonstration, or to discuss your business and technical needs with the Midrange Dynamics team.
Request Info