DataHub
Your Data Foundation — Structured, Searchable, AI-Ready
The AI-Native Data Layer
Data That Works as Hard as Your AI
DataHub isn't just a database—it's the bridge between your information and your AI agents. Design tables visually, define relationships intuitively, and watch as your data becomes instantly searchable, queryable, and actionable by AI without writing a single line of code.
Visual Designer
Drag-and-drop canvas for designing tables, columns, and relationships. See your data architecture come alive.
Smart Relationships
Visual foreign key connections with cascade options. Your data stays connected and consistent automatically.
Semantic Search
PGVector embeddings turn your data into a knowledge base. Search by meaning, not just keywords.
AI Auto-Tools
Every table generates CRUD and search tools automatically. Your agents can access data instantly.
Data Governance Built-In
Table Prefix System
Organize your data lifecycle with standard prefixes that control visibility, AI access, and governance rules. From raw ingestion to curated analytics—every table has its place.
Data Ingestion Layer
Capture raw data from external sources and stage it for transformation. Protected from direct AI modification to preserve data integrity.
- raw_ — Unprocessed external data (AI: Read Only)
- stg_ — Cleaned and validated staging (AI: Read Only)
- sys_ — Internal configuration (Admin Only)
Business Data Layer
Curated, business-ready data that AI agents can read and write. Dimensions for lookups, facts for transactions, curated for operations.
- cur_ — Curated operational data (AI: Read/Write)
- dim_ — Reference and lookup tables (AI: Read Only)
- fct_ — Transactional fact tables (AI: Read/Write)
Rich Column Types
15+ data types with intelligent defaults and validation—powered by PostgreSQL
Text & Numbers
String, Text, Integer, BigInt, Decimal—all PostgreSQL-native types with automatic type coercion.
Dates & Times
Date, DateTime, Time, and Timestamp with timezone support. Perfect for scheduling and audit trails.
Structured Data
JSONB for nested objects, UUID for unique identifiers, Boolean for flags and toggles.
Vector Embeddings
Native VECTOR(1536) columns for RAG embeddings. Semantic search built into your schema.
Enterprise Data Types. Zero Compromises.
From simple strings to complex JSONB documents to vector embeddings—DataHub supports every data type your enterprise needs. Built on PostgreSQL for reliability, enhanced for AI accessibility.
Connected Data Architecture
Visual Relationships
Draw connections between tables on the canvas and DataHub handles the rest—foreign keys, cascade rules, and AI-aware JOINs that make your data work together seamlessly.
One-to-Many
Parent to children relationships—Customer to Orders, Project to Tasks, Invoice to Line Items.
Many-to-One
Reference tables for lookups—Orders reference Customers, Tasks reference Status codes.
Self-Reference
Hierarchical data within a single table—Categories, Org Charts, Threaded Comments.
Cascade Rules
CASCADE, SET NULL, RESTRICT—control what happens when parent records are deleted.
Your Data. Connected. Protected.
AI tools automatically understand relationships. Query a customer and get their orders. Delete a project and cascade to tasks. Your data stays connected and consistent without you writing a single JOIN statement.
RAG-Powered Semantic Search
Turn your data into a knowledge base with PGVector embeddings
Multi-Provider Embeddings
Choose xAI, OpenAI, or local models for embeddings. Use the right model for your data and budget.
Auto-Sync
Embeddings update automatically when data changes. No manual re-indexing, always current results.
Agent Search Tools
Each RAG-enabled table generates a datahub_search tool. Agents find information by meaning.
REST API
Full semantic search API for custom integrations. Build search experiences beyond AI agents.
Search by Meaning. Find What Matters.
Stop building keyword indexes. Enable RAG on your tables and your AI agents can find "customers who complained about shipping" or "products similar to X" without exact matches. Semantic understanding built into your data layer.
Enterprise Security
Multi-Tenant Data Isolation
Every tenant's data is completely isolated at the database level. Row-level security, encrypted storage, and audit logging ensure your data stays protected and compliant.
Row-Level Security
PostgreSQL RLS ensures tenants only see their own data. Isolation at the database engine level.
Encrypted Storage
Data encrypted at rest with AES-256. Sensitive columns can have additional encryption layers.
Backup & Recovery
Automated backups with point-in-time recovery. Your data is protected against loss.
Data Import
Import from CSV, Excel, JSON, or connect to external databases. Bring your data home.
Ready to Build Your AI-Ready Data Foundation?
Design your schema visually, enable semantic search, and give your AI agents instant access to your enterprise data. Start building with DataHub today.