IBMi Knowledge System

From the past to the future. With intelligence.

The AI platform that transforms decades of IBM i code into structured, analyzable knowledge ready for modernization.

The challenge

Legacy code is hidden treasure

Millions of RPG, COBOL, and CL lines power critical processes for thousands of companies. But knowledge is fragmented, programmers retire, and documentation doesn't exist.

🕑

Disappearing knowledge

Veteran programmers leave taking decades of undocumented know-how with them.

🔬

Impossible analysis

Understanding dependencies across thousands of programs, files, and procedures requires weeks of manual work.

🚀

Modernizing blind

Planning a migration without knowing exactly what the code does is risky and expensive.

The platform

Everything you need to understand your IBM i

From automated extraction to AI documentation, IBMi KS is the complete platform for reverse engineering and modernization.

🌐

Knowledge Graph

Automatically builds a Knowledge Graph with 30+ node types: programs, files, fields, procedures, triggers, views, data areas, job queues, journals and their relationships. Field-level cross-reference with McCabe and Halstead complexity metrics.

🔀

Field Lineage

Traces data flow field by field, from source to destination. Export to CSV, JSON, and GraphML for visualization tools.

🤖

AI Chat & Summary

Query code in natural language. AI generates summaries, pseudocode and criticality analysis. Ask to generate Excel, Word or Markdown files with graph data. Enriched with 24,000+ indexed references from official IBM i documentation.

🧠

Use Your Own AI

Bring your own AI provider: Anthropic, OpenRouter, AWS Bedrock, Google Vertex, Azure OpenAI, or local Ollama. Zero vendor lock-in.

🛠

Replatforming Assessment

Automatic complexity scoring, migration proposals with target framework, effort estimation, and risk analysis.

📊

Analytics & Dead Code

Dead code detection, risk scoring with McCabe/Halstead metrics, complete impact analysis (depth 10, no truncation), PII detection, batch/online/report classification, and TCO cost estimation.

🔧

Developer Tools

Call Flow Visualizer, Variable Tracker, SQL Explorer, Indicator Map, Job Flow Diagram, and Cross-Reference Explorer.

📄

Dynamic Reports

Generate professional HTML reports exportable for audits, handovers, and management presentations.

System Integration

Import 30+ IBM i system data sections: users, PTFs, job queues, storage, network, database, TCP/IP services.

🔒

Enterprise Security

RBAC with 12 granular permissions, menu and tab-level access, automated backups to 5 destinations (local, S3, Azure, GCP, Google Drive) with configurable retention policy.

📝

AI Agent Notes

The AI automatically records observations, findings, recommendations and warnings during analysis. An intelligent notepad that builds incremental knowledge visible to the whole team.

The journey: from past to future

IBMi KS transforms your AS/400 heritage into a strategic asset for modernization

1988 — TODAY

The AS/400 heritage

Decades of business logic encoded in RPG, COBOL, and CL. Critical processes running 30+ years whose documentation has been lost.

PHASE 1 — EXTRACTION

Automated discovery

IBMi KS automatically extracts sources, metadata, dependencies, and data structures directly from IBM i systems or X-Analysis exports.

PHASE 2 — ANALYSIS

Knowledge Graph & AI

The Knowledge Graph reconstructs the entire relationship map with field-level cross-reference. AI generates documentation, summaries, complexity metrics and analysis for every object.

PHASE 3 — INSIGHT

Analytics & Assessment

Dead code detection, risk scoring, impact analysis, cost estimation. Everything you need for informed decisions.

PHASE 4 — FUTURE

Guided modernization

Replatforming assessment with architectural proposals, target frameworks, effort estimates, and progressive wave migration plans.

PHASE 5 — AI COLLABORATION

Your AI IDE knows your AS/400

The MCP server connects Claude Code, Cursor, Copilot and IBM i Bob directly to the Knowledge Graph. The AI analyzes, converts and documents with 17 tools, 20 enterprise skills and a notepad where it records every finding — building incremental knowledge accessible to the entire team.

Native integration

Powered by X-Analysis

IBMi KS natively integrates with X-Analysis to import cross-references, dependencies, and metadata extracted directly from IBM i systems. Two complementary tools for complete coverage.

  • Direct import of cross-references and dependencies from X-Analysis exports
  • AI enrichment of the object catalog with automated summaries and documentation
  • End-to-end field lineage combining deterministic parsing and Knowledge Graph
  • Works standalone too — X-Analysis is not a prerequisite
INTEGRATION FLOW
X-Analysis IBMi KS AI Insight
30+
Node types in the graph
6
Supported LLM providers
100K+
Analyzable objects
50+
REST API endpoints

Sized for your reality

From 5,000 to over 100,000 objects — IBMi KS scales with Docker on any infrastructure.

Small — POC/Dev
< 5K objects
vCPU2
RAM4 GB
Disk30 GB
Most common
Medium — SMB
5-20K objects
vCPU4
RAM8 GB
Disk60 GB
Large — Enterprise
20-50K objects
vCPU4
RAM16 GB
Disk120 GB
XLarge — Big Enterprise
50K+ objects
vCPU8
RAM32 GB
Disk250 GB
🧠

Use Your Own AI

No vendor lock-in. Connect the provider you already use or stay fully on-premises with Ollama.

Anthropic Claude
Anthropic Claude — ideal for complex code analysis and documentation generation.
OpenRouter
Unified access to 100+ models: GPT-4o, Gemini, Llama, Mistral and more with a single API.
AWS Bedrock
AWS Bedrock — Amazon-managed models with enterprise security and compliance.
GCP Vertex AI
Google Vertex AI — Gemini and other models within the Google Cloud ecosystem.
Azure OpenAI
Azure OpenAI — GPT-4o on Microsoft infrastructure with enterprise compliance.
Ollama
Ollama — local execution, 100% on-premises. Your data never leaves your network.
Your data, your rules. Each provider is configurable per environment and source code never leaves your infrastructure.
AI IDE Integration

Connect your AI IDE directly to the Knowledge Graph

IBMi KS includes an MCP (Model Context Protocol) server that exposes the entire knowledge graph to AI coding assistants. Claude Code, Cursor, Copilot and any MCP-compatible IDE can query programs, dependencies, source code and AI summaries in real time.

AI without context is guessing

Claude, IBM i Bob, Copilot and Cursor are powerful AI assistants — but without access to YOUR system, they can only give generic answers based on public documentation.

Without IBMi KS

Generic RPG syntax explanations
No knowledge of YOUR programs, files or dependencies
Cannot see call trees, triggers, or data model
Impact analysis is impossible — "I don't have access to your system"
Code conversion without context produces incorrect, untestable code
No awareness of business rules hidden in your codebase
Approximate answers about CL commands, DDS keywords, RPG opcodes

With IBMi KS + MCP

Deep analysis of YOUR specific programs with AI summary
Full call tree, dependencies, file relations, triggers — live
Impact analysis: "Changing CUSTPF impacts 47 programs across 5 levels"
Accurate code conversion with source, data model, and 35 conversion patterns
Business rules extracted from your actual code, not guessed
20 enterprise skills guide the AI like a senior IBM i consultant
24,000+ indexed references from official IBM i documentation

How it works

1

Connect

Add the MCP server URL to your IDE configuration. One line of JSON, zero setup.

2

Ask

Ask your AI assistant about programs, dependencies, impact analysis — in natural language.

3

Convert

Request code conversion with full context: source, call tree, data model, AI summary and conversion skills — all assembled automatically.

📝

AI Notepad: knowledge that accumulates

When an AI agent analyzes a program, it doesn't just produce a temporary output. Every observation, finding, recommendation and warning is saved to a persistent notepad — visible to all users in the frontend.

🤖

The AI writes

While analyzing a program, the agent automatically saves its discoveries: hidden business rules, non-obvious dependencies, migration risks, recurring patterns.

👥

The team reads

Notes are visible on every program's detail page. The junior developer sees what the AI discovered. The senior validates. The manager gets the full picture.

📈

Knowledge grows

Each analysis session adds notes. After 10 sessions, you have detailed documentation no consultant would have ever written. Indexed and searchable.

Works with your favorite AI IDE

🤖

Claude Code

Anthropic's official CLI. Direct MCP integration via stdio or SSE. Ask Claude to analyze, convert or document any AS/400 program with full knowledge graph context.

💻

Cursor

AI-first code editor with native MCP support. Navigate the AS/400 codebase, understand dependencies and generate modern code — all from within the editor.

🤝

GitHub Copilot

GitHub's AI pair programmer. Extend Copilot with deep IBM i knowledge to write accurate conversion code informed by the full dependency graph.

🏭

IBM i Bob

IBM's AI assistant for IBM i. Combine Bob's native platform expertise with IBMi KS's knowledge graph for the most comprehensive analysis possible.

🔌

Any MCP Client

Any tool supporting the Model Context Protocol can connect. The standard is open and growing — future-proof your modernization workflow.

17 MCP tools available

Program info, call tree, impact analysis, file relations, data model
Semantic search, source code, conversion context (with smart chunking)
Read-only SQL queries against the live AS/400 system via ODBC
Agent Notes: the AI records observations, findings and recommendations
20 agent skills + 3 agent profiles (analyst, architect, auditor)

20 enterprise agent skills

Production-ready structured workflows that guide AI agents through every phase of AS/400 modernization — with the same rigor as a senior IBM i consultant.

Analyze — Impact analysis, code understanding, debugging, ODBC real-time queries
Plan — Replatforming planning, data model migration, batch job migration, integration mapping
Build — RPG/CL/DDS conversion patterns, frontend redesign, cloud native services mapping
Verify — Regression testing, data integrity validation, performance benchmarking, security audit
Ship — Cutover planning with hour-by-hour runbook, escalation matrix, rollback procedures

See MCP in action

17
MCP tools
20
Agent skills
6+
Target languages
35
Conversion patterns
24K+
IBM i doc references

Let's start talking

Tell us about your IBM i system and discover how IBMi KS can help with modernization.