Skip to main content

​
Overview

The Search Playground lets you test semantic search across your organization’s content library. Experiment with queries, understand how semantic ranking works, and preview search results before implementing the Search API.

​
Availability & Prerequisites

The Search Playground is available to Enterprise plan customers only. Upgrade to Enterprise to access this feature.
Requirements:
  • Active Enterprise subscription
  • Data Engine access
  • Content uploaded to your content library
You must upload content before using the Search Playground. See Manage Content to add content to your library.

​
Getting Started

Access and use the Search Playground:
  1. Log in to Gloo AI Studio
  2. Click Playground in the main navigation
  3. Select the Search tab
  4. Choose a content library from the Publisher dropdown
  5. Enter a search query in the text input
  6. Press Enter or click Search to see results
Only Publishers where you have access will appear in the dropdown. If no Publishers are visible, verify that content has been uploaded to your organization.
Publisher dropdown showing available content libraries

​
Understanding Search Results

Search results display as cards, each containing: Title: The content item’s headline or name Part Number: Sequential identifier for multi-part content (e.g., chapters, episodes) Source: The publisher or origin of the content Snippet: A preview of relevant text from the content Results are ranked by semantic relevance, not keyword matching. The most contextually relevant content appears first, even if it doesn’t contain your exact search terms.
Semantic search understands meaning and context. A search for “dealing with anxiety” might return content about “managing stress” or “finding peace” because they’re semantically related.
Search results showing cards with titles, sources, and snippets

​
Exploring Key Insights

Some search results include expandable Key Insights sections:
  1. Look for the Key Insights section below search results
  2. Click to expand and view AI-generated summaries
  3. Review insights to understand common themes across results
Key Insights are automatically generated when available. Not all searches will include insights depending on result quality and content type.
Expanded Key Insights panel showing AI-generated summaries Get better results with these best practices: Use Natural Language Write queries as questions or phrases, not just keywords:
  • Good: “How can I improve my prayer life?”
  • Also Good: “tips for consistent prayer”
  • Less Effective: “prayer”
Be Specific Add context to narrow results:
  • “marriage advice for newlyweds” instead of “marriage”
  • “understanding the book of Romans” instead of “Romans”
Test Variations Try different phrasings to see how semantic search interprets your query:
  • “coping with loss” vs. “grief support” vs. “dealing with death”
Reset and Retry Click Reset to clear your search and start fresh with a new query.

​
Transitioning to API

Once you’ve tested search queries in the Playground, implement them via the Search API: Parameter Mapping:
Playground SettingAPI ParameterExample Value
Publisher selectiontenant and collectionOrganization and publisher IDs
Search querysearch"How to pray effectively"
Result countlimit10 (default)
Your Playground experiments translate directly to API requests. Use successful queries to guide your production search implementation.

Search API Guide

Implement semantic search in your application

Manage Content

Upload and organize content in Studio

API Credentials

Generate credentials for API access

Data Engine

Learn about content ingestion and management
Semantic search differs from traditional keyword search by understanding meaning and context: Keyword Search:
  • Matches exact words in your query
  • Ranks by word frequency
  • Misses synonyms and related concepts
Semantic Search:
  • Understands intent and meaning
  • Ranks by contextual relevance
  • Finds related concepts even without exact matches
Example Comparisons:
QueryKeyword Search FindsSemantic Search Finds
”finding purpose”Content with exact phrase “finding purpose”Also: “discovering your calling”, “life meaning”, “God’s plan for you"
"dealing with fear”Only “fear” mentionsAlso: “overcoming anxiety”, “courage”, “trusting God in uncertainty"
"healthy relationships”Exact match onlyAlso: “building strong marriages”, “biblical friendship”, “communication skills”
Think about what users mean, not just what they say. Semantic search finds content that answers the underlying question, even if the wording differs.

​
Frequently Asked Questions

Why can’t I access the Search Playground? Search Playground requires an Enterprise plan with Data Engine access. Free and Pay-as-you-go plans do not include search capabilities. Contact sales to upgrade. Why don’t I see any Publishers in the dropdown? Publishers only appear after content has been uploaded. Upload content via Manage Content or the Content API first, then return to the Search Playground. How many results does the Playground return? The Playground returns up to 10 results by default. The Search API lets you customize the result limit using the limit parameter (up to 100 results per request). Why are my results irrelevant? Try these troubleshooting steps:
  • Make your query more specific with additional context
  • Verify your content library contains relevant material
  • Test different query phrasings
  • Ensure content has been fully processed after upload
What’s the difference between semantic and keyword search? Semantic search understands meaning and context, finding conceptually related content even without exact word matches. Keyword search only matches specific words. Semantic search provides better results for natural language queries. Can I filter results by content type? The Playground shows all results from the selected Publisher. Use the Search API to implement custom filtering by content type, date, author, or other metadata in your application. How do I use search results for RAG (Retrieval Augmented Generation)? Search results include text snippets perfect for RAG implementations. The snippets preview the exact content you’ll receive via the API, which you can pass to the Completions API as context for generating answers.