Skip to content

Indexing Quick Start

Get started with the new Indexing Tab interface in just a few minutes. This guide shows you how to create your first index and start searching your data.

Prerequisites

Before you begin, ensure you have:

  • ✅ PgVector configured - Set up in AI Configuration
  • ✅ Embedding Model selected - Choose your preferred model
  • ✅ Toolkit with indexing support - GitHub, Confluence, Jira, etc.

Quick Steps

1. Access the Indexing Interface

  1. Navigate to Toolkits in the main navigation
  2. Select a toolkit that supports indexing (e.g., GitHub, Confluence, Jira)
  3. Click the "Indexes" tab in the toolkit detail view (enabled when "Index Data" tool is selected and PgVector + embedding model are configured)
  4. The Indexes interface opens with sidebar showing existing indexes

Index Tab

2. Create Your First Index

  1. Click the + Icon (+ Create New Index button) in the Indexes sidebar
  2. Fill required fields in the new index creation form:
    • Index Name (Collection Suffix): Unique name for your index (max 7 characters, e.g., docs, main)
    • Toolkit-specific parameters (varies by toolkit type)
  3. Configure optional settings:
    • Progress Step: Progress reporting interval (0-100)
    • Clean Index: Remove existing data before indexing
    • Chunking Config: Document chunking configuration
  4. Click "Index" button to start processing (button activates when required fields are filled)

3. Start Searching

Once indexing completes:

  1. Select your completed index from the sidebar
  2. Navigate to Run Tab in the center panel
  3. Choose a search tool from the dropdown menu:
    • Search Index: Basic semantic search
    • Stepback Search Index: Advanced search with query decomposition
    • Stepback Summary Index: Search with automatic summarization
  4. Enter your query (e.g., "How do I configure authentication?")
  5. Configure search parameters (optional): Filter, Cut Off, Search Top, etc.
  6. Click Run to execute search and view results in the right panel chat interface

Example: GitHub Repository Index

Here's a complete example for indexing a GitHub repository:

Step 1: Create Index

  • Click + Create New Index in sidebar
  • Index Name: docs (collection suffix)
  • Configure GitHub-specific parameters (repository, branch, etc.)
  • Click Index button

Step 2: Monitor Progress

  • Watch progress bar in center panel
  • Wait for ✅ Completed status (~2-5 minutes depending on size)

Step 3: Search

  • Select completed index from sidebar
  • Navigate to Run tab
  • Choose Search Index from tool dropdown
  • Enter query: "API authentication best practices"
  • Click Run button

🎉 You're ready! Start indexing your data and unlock powerful search capabilities across your content.

Next Steps