100% Open Source Today
⚡ Enterprise Semantic CMDB Tomorrow

The Enterprise Platform for Self-Learning Semantic CMDBs

VisionML.Net is pioneering the transformation of network infrastructure management through Semantic CMDB technology with self-updating knowledge graphs.

Start your journey today with NetIntel-OCR, our free, open-source foundation that extracts network intelligence from documents - the first step toward tomorrow's self-learning enterprise CMDBs.

100% Private Mode
Zero Data Leakage
Compliance Ready
The Semantic CMDB Evolution
Phase 1: Document Intelligence
Available Now
Phase 2: RDF Generation
Roadmap Q4 2025
Phase 3: Self-Learning CMDB
Auto-Updating Graphs
The Hidden Crisis

Your Enterprise Knowledge is Trapped and Dying

10+ years of critical network and security documentation sits dormant in Confluence pages, SharePoint folders, and aging wikis - invisible to modern AI systems and inaccessible when you need it most.

Documentation Graveyards

  • Engineering docs from 2014 still define critical systems
  • Security runbooks scattered across 1000s of Confluence pages
  • Network diagrams in outdated Visio files on SharePoint
  • Operations guides known only by retiring engineers

Tribal Knowledge Crisis

  • 67% of network knowledge exists only in people's heads
  • Average tenure: 2.5 years - knowledge walks out the door
  • 3AM incidents depend on who remembers what
  • $4.5M average cost of knowledge loss per year

AI Can't Help You

  • LLMs can't read your SharePoint or Confluence
  • ChatGPT doesn't know your network architecture
  • Copilots fail without access to your documentation
  • Automation stops at undocumented processes

The Data Speaks for Itself

10+ Years
of documentation debt
80%
of knowledge undiscoverable
45 mins
average search time per incident
$2.5M
annual cost of rework
The Solution

Transform Dead Documentation into Living AI Knowledge

NetIntel-OCR liberates your trapped knowledge, making it LLM-ingestible and available through MCP (Model Context Protocol) for conversational AI and agentic workflow automation.

From Documentation Graveyards to AI-Ready Knowledge

1

Extract Everything

Scan Confluence, SharePoint, wikis, PDFs - pull out 10+ years of dormant knowledge

2

Structure for AI

Convert to vectors, knowledge graphs, and LLM-digestible formats

3

Serve via MCP

Model Context Protocol makes knowledge available to any LLM or AI agent

4

Enable Intelligence

Power conversational AI, automated workflows, and intelligent operations

MCP: The Bridge to AI

# MCP Service Endpoint
POST /mcp/v1/context
{
"query": "network topology for datacenter-1",
"sources": ["confluence", "sharepoint", "runbooks"],
"format": "knowledge-graph"
}
Works with Private Inferencing vLLM, OLLAMA, MLX models
Private LoRA Fine-Tuning using NVIDIA EdgeAI Hardware
Enables GitHub Copilot for infrastructure
Powers agentic automation workflows
Maintains security boundaries
90%
Knowledge Recovery
From 10+ years of documentation
5 mins
To Find Answers
Down from 45 minutes
100%
AI-Accessible
Via MCP protocol
The VisionML.Net Platform

VisionML.Net: Your Path to Semantic CMDB

VisionML.Net is an enterprise platform designed to transform how organizations manage network infrastructure through Semantic CMDB technology. We believe the journey to intelligent, self-updating knowledge graphs should be accessible to everyone.

Start Today with NetIntel-OCR

NetIntel-OCR is our open-source foundation - a fully functional, production-ready tool that extracts network diagrams, tables, and text from PDFs. It's free, Apache 2.0 licensed, and delivers immediate value while serving as your on-ramp to the VisionML.Net platform.

Enterprise Security First

100% Private Mode Operation

Your data never leaves your infrastructure. Built for the most demanding regulatory and compliance requirements.

Air-Gapped Deployment

Operates completely within your firewall. No vectors, knowledge graphs, or data ever leave your enterprise boundaries. Perfect for classified environments.

EdgeAI GPU Processing

Leverage your on-premise GPU resources for AI processing. No cloud dependencies. Supports NVIDIA, AMD, and Intel GPUs for maximum performance.

Compliance Certified

Designed for GDPR, HIPAA, SOC 2, ISO 27001, and government compliance. Full audit trails and encryption at rest and in transit.

Cloud-Native Kubernetes Architecture

Scalability Without Compromise

  • Deploy on your private Kubernetes clusters - on-premise or private cloud
  • Horizontal scaling with distributed processing across edge nodes
  • GPU scheduling for AI workloads with Kubernetes device plugins
  • Multi-tenancy support with namespace isolation

Zero Trust Security Model

  • All data encrypted with customer-managed keys (BYOK)
  • Network policies enforce strict pod-to-pod communication
  • Service mesh with mTLS for all internal communications
  • RBAC and admission controllers for access control

Compliance & Certification Ready

NIST CSF
NIST 800-53
GDPR Compliant
HIPAA Ready
FedRAMP Compatible
SOC 2 Type II
ISO 27001

The Semantic CMDB Promise

From document extraction today to intelligent operations tomorrow

70%
Faster Incident Resolution
85%
Faster Change Planning
95%
Less Audit Time
10x
Better Planning Intelligence
Open Source Foundation

From Open Source Today to
Self-Learning Enterprise CMDB

Current State: NetIntel-OCR is a fully functional, production-ready open-source tool delivering immediate value through intelligent document processing. Extract network diagrams, tables, and create searchable vector databases today.

Future Vision: Evolution into an enterprise-grade Semantic CMDB platform featuring self-updating knowledge graphs that automatically learn from network changes, predict cascading impacts, and enable SPARQL-based intelligence queries.

Apache 2.0 License
100% On-Premise
Community Driven

Self-Updating Knowledge Graphs

1
Continuous Learning
Automatically detects and incorporates network changes into the knowledge graph
2
Relationship Discovery
Identifies new connections and dependencies as infrastructure evolves
3
Anomaly Detection
Learns normal patterns and alerts on configuration drift
4
Predictive Intelligence
Forecasts impact of changes based on historical patterns
5
CMDB Integration
Auto-updates ServiceNOW, SolarWinds, Device42 with discovered changes

Building Blocks of Semantic Intelligence

Today's document processing capabilities become tomorrow's knowledge graph foundation

Predictable AI OPEX Costs

Eliminate unpredictable cloud AI spending with local inferencing on existing infrastructure.

  • Fixed costs with on-prem hardware
  • No per-token cloud charges
  • Leverage existing CPU infrastructure
  • Scale without OPEX surprises

Local-First AI Inferencing

NVIDIA and Apple MLX friendly hardware for dynamic business environments without cloud dependencies.

  • Apple MLX optimization
  • NVIDIA CUDA support
  • CPU-first, GPU when needed
  • Hybrid on-prem/cloud flexibility

Private Mode Security

100% private operation - no vectors, data, or knowledge ever leaves your firewall.

  • Air-gapped deployment
  • GDPR/HIPAA compliant
  • Customer-managed keys
  • Zero trust architecture

AI-Powered Extraction

Automatic detection and extraction of network diagrams, tables, and text using advanced AI models.

  • Multi-model support
  • Vision-language models
  • Ollama/vLLM/LMCache integration

EdgeAI GPU Processing

Leverage your on-premise GPU infrastructure for AI workloads. No cloud GPU costs or data transfers.

  • NVIDIA A100/H100 support
  • AMD MI300 compatible
  • Intel Gaudi acceleration
  • Multi-GPU scaling

Milvus Vector Database

Enterprise-scale vector search with Milvus integration for lightning-fast semantic queries.

  • Sub-100ms queries
  • Distributed architecture
  • 4096-dimensional embeddings

Network Diagram Processing

Convert visual network diagrams to structured Mermaid.js format with component detection.

  • Topology mapping
  • Component recognition
  • Connection tracing

C++ Performance Core

50-100x performance boost with AVX2 SIMD and OpenMP parallelization for deduplication.

  • Three-level deduplication
  • Zero-compilation install
  • SimHash fuzzy matching

Cloud-Native Architecture

Kubernetes-native deployment with EdgeAI GPU acceleration for private cloud scale.

  • Private Kubernetes clusters
  • EdgeAI GPU scheduling
  • Horizontal auto-scaling
  • Zero cloud dependency
MCP Integration

Model Context Protocol (MCP) Service

Transform your 10+ years of trapped documentation into LLM-ingestible knowledge. MCP makes your enterprise knowledge available to any AI system or automation workflow.

Private Conversational AI

Enable private LLMs via vLLM, OLLAMA, and MLX to answer questions about your network infrastructure. Fine-tune models with LoRA on NVIDIA EdgeAI hardware.

Agentic Workflows

Power autonomous agents that can navigate your documentation, understand dependencies, and execute complex network operations.

Knowledge Sync

Continuously extract from Confluence, SharePoint, and wikis. Keep your AI knowledge base current with automated synchronization.

MCP in Action

# Query your enterprise knowledge
$ netintel-ocr --mcp-query \
"Show me the firewall rules for DMZ"
# Returns from 10 years of docs:
✓ Confluence: 2019 Security Runbook
✓ SharePoint: 2021 Network Diagram
✓ Wiki: 2023 Change Request #1247
# Private LLM with your data
$ export MCP_ENDPOINT=localhost:8001
$ vllm serve --model llama3-finetuned
> "What's the VLAN config for Building A?"
# Private LLM responds with YOUR data:
"Based on your documentation from
2018-2024, Building A uses VLANs
100-150 for production..."

Private Cloud-Native Architecture

Deploy on your infrastructure - from edge nodes to private Kubernetes clusters

100% Private Kubernetes Native EdgeAI Ready

CLI Entry Points

--init

Project initialization

--batch-ingest

Parallel processing

--query

Vector search

--api / --mcp

Server modes

Processing Pipeline

Text Extraction

nanonets-ocr model

Network Detection

qwen2.5vl model

Table Extraction

pdfplumber library

Vector Generation

Milvus/LanceDB

Storage & Integration

Milvus

Vector database

MinIO/S3

Object storage

etcd

Metadata store

REST API

Integration endpoint

Private Deployment Options

Scale within your firewall - from edge to enterprise

All deployments run 100% on your infrastructure

Minimal

Personal use

  • 1-5 users
  • <10 docs/day
  • 5 min setup
  • 2GB RAM

Small

Team use

  • 5-20 users
  • 10-50 docs/day
  • 15 min setup
  • 4GB RAM

Medium

Department

  • 20-100 users
  • 50-200 docs/day
  • 30 min setup
  • 8GB RAM

Enterprise

Organization-wide

  • 100+ users
  • 200+ docs/day
  • 1 hour setup
  • 16GB+ RAM

Quick Start in 5 Minutes (100% Private)

All processing happens on your infrastructure. No data ever leaves your network.

Install & Setup

# Install NetIntel-OCR
pip install netintel-ocr
# Install high-performance Ollama/vLLM/LMCache
curl -fsSL https://ollama.com/install.sh | sh
ollama serve
# Pull required models
ollama pull nanonets-ocr-s:latest
ollama pull qwen2.5vl:latest

Process Your First PDF

# Process a document
netintel-ocr document.pdf
# With Docker support
netintel-ocr --init --deployment-scale minimal
cd netintel-ocr/docker
docker-compose -f docker-compose.minimal.yml up

The Semantic CMDB Revolution

From static documentation to intelligent, self-aware network knowledge

Today: Document Intelligence

Extract Network Diagrams
Convert PDFs to structured Mermaid.js format
Structure Configuration Data
Tables become queryable JSON
Vector Search
Find documents 70% faster

Tomorrow: Semantic CMDB

Auto-Update Existing CMDBs
Sync with ServiceNOW, SolarWinds, Device42
Understand Relationships
RDF triples express network connections
Predict Impacts
SPARQL queries reveal dependencies
Self-Updating Intelligence
Living knowledge that evolves with your network

The Evolution Path

Phase 1: Available Now

Document extraction, vector search, structured data

Phase 2: Q4 2025

RDF generation, knowledge graphs, basic queries

Phase 3: Q1 2026

Full semantic intelligence, automated operations

Industry Focus

Purpose-Built for Critical Infrastructure

VisionML.Net is designed specifically for industries where network documentation accuracy, security, and compliance are mission-critical.

Telecommunications

5G, SD-WAN, Network Slicing

Complex Network Topologies
Handle multi-vendor, multi-technology environments
Service Assurance
Track SLAs and network performance documentation
Regulatory Compliance
Meet FCC, GDPR, and regional requirements
Key Use Case: Extract and maintain 5G network slice configurations from 10+ years of vendor documentation

Managed Security

MSSPs, SOCs, MDR Providers

Multi-Tenant Architecture
Isolate customer documentation and configurations
Security Runbooks
Convert incident response docs to AI-accessible knowledge
Compliance Automation
NIST CSF, NIST 800-53, SOC 2 documentation
Key Use Case: Transform thousands of security runbooks into AI-powered incident response automation

Healthcare

Hospitals, Health Systems, HIEs

HIPAA Compliance
Maintain PHI security with air-gapped deployment
Medical Device Networks
Document IoMT and clinical system architectures
Disaster Recovery
Maintain critical infrastructure documentation
Key Use Case: Extract medical device network topologies from vendor PDFs for security assessments

Why These Industries Choose VisionML.Net

24/7
Critical Operations
100%
Data Sovereignty
Zero
Cloud Dependency
Full
Audit Trail
Built by Experts

Created by Network Security Professionals

VisionML.Net is built by proven network security professionals with 15+ years of open-source experience, DoD expertise, and collaboration with IFML Research Labs at UT Austin.

CNCF Contributors

Core contributors to Cloud Native Computing Foundation projects including Kubernetes networking and service mesh architectures.

  • Kubernetes CNI plugins
  • Envoy proxy contributions
  • Istio service mesh

Security Expertise

15+ years securing critical infrastructure for Fortune 500 companies and government agencies worldwide.

  • Zero-trust architecture design
  • NIST framework implementation
  • Security automation tools

Open Source Leaders

Maintainers and contributors to major open-source networking and security projects used by millions.

  • 10,000+ GitHub contributions
  • 50+ open-source projects
  • Apache & CNCF members

Our Distinguished Background

Open Source Innovation

Pioneered CNI (Container Network Interface) standards
Built eBPF-based network observability tools
Contributed to Calico, Cilium, and Flannel
Developed Falco runtime security rules

DoD & Government Experience

DoD classified network architecture
Zero-trust implementation for defense systems
DISA STIG compliance automation
Critical infrastructure protection

IFML Research Labs - UT Austin Partnership

Collaborating with the Institute for Foundations of Machine Learning at The University of Texas at Austin to advance semantic CMDB technology.

Knowledge graph research
Self-learning algorithms
Network topology inference
Automated reasoning systems

"Combining 15+ years of open-source expertise, DoD security experience, and cutting-edge research from UT Austin's IFML Labs to solve network documentation at scale."

- The VisionML.Net Team

Enterprise Integration Roadmap

Seamless Integration with Your Existing CMDB

VisionML.Net doesn't replace your CMDB - it makes it intelligent. Our self-updating knowledge graphs will automatically sync discovered changes to your existing systems.

ServiceNOW

Bi-directional sync with ServiceNOW CMDB. Auto-update CIs, relationships, and dependencies discovered from network documentation.

  • Configuration Item updates
  • Relationship mapping
  • Change detection alerts
  • Impact analysis integration
Available: Q4 2025

SolarWinds

Enhance SolarWinds Orion with semantic understanding. Auto-discover and update network topology from documentation.

  • Network topology sync
  • Device inventory updates
  • Dependency mapping
  • Alert correlation
Available: Q1 2026

Device42

Complete Device42 integration for hybrid infrastructure management. Keep your asset database current with discovered changes.

  • Asset discovery sync
  • Application dependency updates
  • IP address management
  • Rack layout validation
Available: Q1 2026

How Auto-Update Works

1. Process Docs

NetIntel-OCR extracts network data from PDFs

2. Build Knowledge

Create semantic graphs with relationships

3. Detect Changes

Identify updates and new configurations

4. Sync CMDBs

Auto-update ServiceNOW, SolarWinds, Device42

Start Today, Transform Tomorrow

The practical path to revolutionary Semantic CMDB capabilities

Immediate Value (Available Now)

70% Faster Documentation Search
Find network information instantly with vector search
Structured Network Diagrams
Legacy PDFs become queryable Mermaid.js
Enterprise-Scale Processing
Handle thousands of documents with complete data sovereignty

The Semantic CMDB Future

What Semantic CMDBs Enable:

"What services fail if this router goes down?"
"Show me all attack paths to our databases"
"Where will we hit capacity limits next?"
Your phased journey:
Document Processing → RDF Generation → Full Intelligence

Open Source Today, Enterprise Tomorrow

Start free with our open-source foundation, scale to enterprise Semantic CMDB

Open Source Core

Free

Forever - Available Now

  • Full document processing
  • Network diagram extraction
  • Vector search capabilities
  • Apache 2.0 License
  • Community support
  • 100% on-premise
PRIVATE ALPHA

Semantic CMDB Pro

Coming Soon

Q3 2025

  • Everything in Open Source
  • RDF triple generation
  • Knowledge graph builder
  • SPARQL query interface
  • Impact analysis
  • Priority support

Enterprise Semantic

Custom

2026 Roadmap

  • Everything in Pro
  • Self-updating knowledge graphs
  • ML-powered learning
  • Predictive analytics
  • Anomaly detection
  • White-glove support

Begin Your Semantic CMDB Transformation

Don't wait for the future of network intelligence—build it today. NetIntel-OCR delivers immediate value while laying the foundation for tomorrow's Semantic CMDB revolution.

Start with document extraction, evolve to knowledge graphs, transform to intelligent operations.