Skip to main content
POST
/
ai
/
data
/
v1
/
search
Search publisher data
curl --request POST \
  --url https://platform.ai.gloo.com/ai/data/v1/search \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "query": "How can I find true happiness?",
  "collection": "GlooProd",
  "tenant": "TestOrg",
  "certainty": 0.5,
  "limit": 1
}
'
{
  "data": [
    {
      "uuid": "12345678-abcd-1234-abcd-1234567890ab",
      "metadata": {
        "creation_time": "2025-09-02T23:49:49.338000Z",
        "last_update_time": "2025-09-02T23:50:36.348000Z",
        "distance": 0.39595216512680054,
        "certainty": 0.8020238876342773,
        "score": 0,
        "explain_score": "",
        "is_consistent": false,
        "rerank_score": null
      },
      "properties": {
        "item_title": "Finding True Happiness",
        "publisher": "Test Org",
        "publication_date": "2025-09-02",
        "snippet": "# Finding True Happiness: A Christian Perspective",
        "type": "Article",
        "item_id": "12345678-abcd-1234-abcd-1234567890aa"
      },
      "references": null,
      "vector": {},
      "collection": "GlooProd"
    }
  ],
  "intent": 1
}

Documentation Index

Fetch the complete documentation index at: https://docs.gloo.com/llms.txt

Use this file to discover all available pages before exploring further.

This is a semantic search endpoint. It searches by meaning, not by exact keyword or prefix matching. Write queries as natural language questions or phrases for best results.

​
Before You Start

This endpoint uses AI-powered semantic search to find content based on the meaning of your query — not exact text matching. For example, a query like “secrets to a happy marriage” will return content about “rules for keeping a marriage healthy,” even though none of the exact words overlap. This is not a keyword or prefix search. Queries like "happ" or "happiness AND marriage" will not behave as expected. Use complete, natural-language phrases instead.

Search Developer Guide

Understand how semantic search works, how to tune the certainty threshold, and how to combine search with RAG.

Search Tutorial

Working code examples in 6 languages — JavaScript, TypeScript, Python, PHP, Go, and Java.

Error Reference: Search

Common errors: invalid collection, missing or invalid tenant, certainty threshold issues, empty query string.

​
Tips for Effective Queries

  • Use natural language — “How can I find true happiness?” works better than “happiness find true”
  • Adjust certainty — The default threshold is 0.75, which is strict. If you’re getting few or no results, try lowering it to 0.5
  • Results are ranked by semantic distance — not keyword frequency. The certainty and distance fields in the response indicate how closely each result matches your query’s meaning

Authorizations

​
Authorization
string
header
required

Bearer authentication header of the form Bearer <token>, where <token> is your auth token.

Body

application/json
​
query
string
required

Search query.

Minimum string length: 1
​
collection
string
required

Collection to search in.

Minimum string length: 1
​
tenant
string

Optional tenant to search in. Required only when the selected collection is multi-tenant; invalid when the selected collection is not multi-tenant.

​
certainty
number
default:0.75

The minimum certainty threshold for snippets.

Required range: 0 <= x <= 1
​
limit
integer
default:10

Maximum number of results to return.

Required range: x >= 1

Response

Successful response

​
data
object[]
required
​
intent
integer
required