Skip to main content

Knowledge Bases

Store, organize, and search through your data with structured tables, semantic search, and powerful query capabilities.

Platform Requirement

Knowledge Bases are only available in the Timbal Platform.

You must create your Knowledge Base from the Timbal Platform, and a valid Timbal API token is required to access it.

Throughout this documentation, any references to org ID and kb ID refer to the identifiers provided by the Timbal Platform for your organization and knowledge base.

What is a Knowledge Base?​

Think of a Knowledge Base as a smart database that combines the best of both worlds:

Traditional Database Capabilities​

  • Store structured data in organized tables (like spreadsheets)
  • Query data with precise SQL commands
  • Maintain data relationships and integrity
  • Import data from files or add it programmatically

AI-Powered Intelligence​

  • Search using natural language ("Find documents about API integration")
  • Understand context and meaning, not just exact words
  • Find similar content automatically
  • Learn from your data patterns

The Three Building Blocks​

Knowledge Bases are built on three core components that work together:

1. Tables - Your Data Foundation​

Think of tables as organized filing cabinets. Each table is like a cabinet with labeled drawers (columns) where you store specific types of information.

Example: A "Products" table might have drawers for:

  • Product name
  • Description
  • Price
  • Category
  • Stock quantity

2. Embeddings - The AI Brain​

Embeddings are like giving your data a "brain" that understands meaning. They convert text into mathematical representations that capture the essence of what the text is about.

How it works:

  • "Laptop computer" and "portable PC" get similar mathematical representations
  • "Customer support" and "help desk" are recognized as related concepts
  • Natural language queries find relevant content even without exact matches

3. Indexes - The Speed Boosters​

Indexes are like the index at the back of a book - they help you find information quickly without reading every page. They create fast pathways to your data.

Types of speed boosts:

  • B-tree indexes: Fast lookups for names, dates, numbers
  • Hash indexes: Lightning-fast exact matches
  • GIN indexes: Powerful text search capabilities
  • Composite indexes: Optimized for complex queries

How Knowledge Bases Transform Your Data​

Before Knowledge Bases​

User: "I need help with password reset"
System: ❌ No results found (exact match only)

With Knowledge Bases​

User: "I forgot my login credentials"
System: ✅ Found: "How to reset your password" 
      ✅ Found: "Account recovery process"
      ✅ Found: "Troubleshooting login issues"

The Knowledge Base Workflow​

1

Design Your Data Structure

Plan what information you want to store and how it should be organized.

2

Create Tables

Set up the structure for your data with appropriate columns and data types.

3

Import Your Data

Add your information from files, databases, or programmatically.

4

Enable AI Search

Create embeddings on text columns to enable semantic search capabilities.

5

Optimize Performance

Add indexes to ensure fast queries as your data grows.

6

Search and Discover

Getting Started​

Ready to build your first Knowledge Base? Start with these guides:

  • Tables - Learn how to create and organize your data
  • Embeddings - Enable AI-powered semantic search
  • Indexes - Optimize performance for fast queries