AutoWarehouse
Platform Overview

Built for Enterprise Data

Multi-tenant, AI-powered, self-hosted. One platform from source ingestion to business intelligence.

Platform Capabilities

Enterprise Web Platform
Multi-tenant UI with role-based dashboards and real-time collaboration
AI Agent Engine
Multi-agent orchestration for discovery, mapping, and documentation
LLM Gateway
Local-first LLM integration β€” Ollama, vLLM, LiteLLM, or any compatible runtime
Data Profiling Engine
Automated database discovery, column profiling, anomaly detection
ETL Code Generator
Deterministic, plugin-based exporters for Airflow, dbt, Snowflake, Spark, PL/SQL
Data Quality Engine
Null analysis, cardinality checks, anomaly flags, completeness scoring
Metadata Store
PostgreSQL with vector extensions for semantic search and audit trails
Container-Native
Docker Compose and Kubernetes-ready with Helm charts

Security & Compliance

πŸ”
AES-256-GCM
Encryption for all credentials
πŸ”‘
OIDC-First Auth
Keycloak, Auth0, LDAP support
πŸ›‘οΈ
RBAC
30+ granular permissions
πŸ“‹
Audit Logging
Full traceability for every AI decision and data operation

Complete Data Isolation

AutoWarehouse runs entirely within your network perimeter. No cloud dependency, no external API calls, no data leaves your environment β€” ever.

βœ“ Air-Gapped Deployment

The entire platform β€” application, AI engine, LLM runtime, and database β€” runs on your infrastructure without any internet connectivity requirement. Deploy in fully isolated, air-gapped networks that meet the strictest security requirements of banking, defense, and government environments.

βœ“ Local LLM β€” No Cloud AI

AI capabilities run on self-hosted LLM runtimes (Ollama, vLLM, LiteLLM). Your source schemas, column names, business rules, and mapping logic are never sent to any external service. The AI engine lives inside your firewall.

βœ“ Zero Data Exfiltration

AutoWarehouse connects to your source databases read-only for profiling and discovery. Generated ETL code runs in your own pipeline tools (Airflow, dbt, Spark). No data is copied, cached, or transmitted outside your environment.

βœ“ Compliance Ready

Designed for regulated industries: KVKK, GDPR, PCI-DSS, SOC 2 β€” the platform's architecture ensures data residency compliance by design. Full audit trails for every AI decision and data operation.

For enterprise security teams: AutoWarehouse requires no outbound network access. It can be deployed in environments with no internet connectivity. All dependencies β€” including LLM weights β€” are bundled in the container images.

Intellica Ecosystem

AutoWarehouse integrates with and builds upon Intellica's proven product family

Data Models
IFDM Β· ITDM Β· HRDM

Industry-standard warehouse models for Finance, Telecom, and HR β€” refined across dozens of deployments

Quality
ICC

Central data consistency management β€” reconciliation rules, quality gates, and drift monitoring

NL Analytics
Talk To

Natural language data interaction β€” ask questions, get SQL queries and visualizations instantly

Orchestration
Orqenta

Workflow orchestration and automation for regulated environments and complex data pipelines

Deploy Your Way

Self-hosted on your infrastructure. No cloud lock-in. Choose the topology that fits your scale.

Development & POC

Single Machine

Docker Compose on a single VM or laptop. Ideal for evaluation, development, and proof-of-concept deployments.

$ docker compose up -d
βœ“ autowarehouse-db β€” metadata & vector store
βœ“ autowarehouse-web β€” application server
βœ“ autowarehouse-worker β€” ETL processing engine
βœ“ autowarehouse-llm β€” local LLM runtime
Min: 4 vCPU Β· 16 GB RAM Β· 100 GB SSD
Production

Multi-Container

Isolated services with dedicated resources, external database, and self-hosted LLM runtime. Production-ready with monitoring.

β—† External PostgreSQL (RDS, Cloud SQL, on-prem)
β—† Horizontal worker scaling (N ETL workers)
β—† Local LLM: Ollama, vLLM, or LiteLLM
β—† Reverse proxy (Nginx / Traefik) + TLS
β—† Log aggregation & health checks
Rec: 8+ vCPU Β· 32+ GB RAM Β· HA storage
Enterprise Scale

Kubernetes

Helm charts for full Kubernetes orchestration. Auto-scaling, multi-tenant isolation, and enterprise-grade observability.

β—† Helm chart with configurable values
β—† HPA auto-scaling (web + workers)
β—† Namespace-level tenant isolation
β—† Prometheus + Grafana monitoring
β—† Backup & disaster recovery
EKS Β· AKS Β· GKE Β· On-prem K8s
Platform Architecture
Presentation
Web Application
Dashboards Β· Data Studio Β· Auth
AI Engine
Multi-Agent Orchestration
Discovery Β· Mapping Β· NL-SQL
ETL Runtime
Worker Pool + Code Generator
Execution Β· Profiling Β· Export
Data Layer
Metadata & Vector Store
Catalog Β· Embeddings Β· Audit
πŸ”“
Data Sovereignty
Your data never leaves your infrastructure. No external API calls required β€” run LLMs locally with Ollama, vLLM, or LiteLLM.
πŸ“ˆ
Elastic Scaling
Scale ETL workers independently from the web tier. Process 10 tables or 10,000 β€” the architecture adapts to your workload.
πŸ”„
Zero Downtime Updates
Rolling deployments with database migrations. Update the platform without interrupting running ETL jobs or active sessions.