# Hybrid Search

> :test\_tube: **Beta Function :** This function is currently in Beta. Future updates will be documented here.

## Overview

The **Vector Hybrid Search** feature enhances the search experience by combining traditional keyword-based search with vector-based semantic search. This approach provides more comprehensive and relevant results by leveraging both exact keyword matching and understanding the meaning behind queries.

**Key Capabilities:**

* **Comprehensive Results**: Combines keyword-based precision with vector-based semantic understanding.
* **Adjustable Weighting**: Control the balance between keyword and vector search results.
* **Minimum Similarity Threshold**: Set quality standards for vector search results.
* **API Parameter Override**: Fine-tune vector search behavior on a per-request basis.

By combining both search approaches, Hybrid Search delivers results that are both precise and contextually relevant, helping users find what they're looking for even when their terminology differs from product descriptions.

**Note:** This feature leverages vector-based language understanding technology but does **not** rely on general-purpose AI or large language models (LLMs). Your data stays on our servers, locally.

## Vector Search Settings

The **Vector Search Settings Page** allows you to configure how keyword and vector search results are combined. With this feature, you can enable **hybrid search mode** and control the weight and quality of vector search results.

**Availability**

This settings page is **only visible** if the **Vector Search module is enabled** for your account.

![Vector Search Hybrid Mode](/files/D7qOCYP9qi2ZuYnXgKIq)

### Settings

**Vector Search Mode**

* **Description**: Determines how vector search is used in relation to keyword search.
* **Options**:
  * **Only as fallback**: Vector search will only be used when keyword search returns no results.
  * **Hybrid**: Keyword and vector search results will be merged.
* **Default**: **Only as fallback**

**Hybrid Weight**

* **Description**: Controls the weight of vector search results when in hybrid mode.
* **Range**: 0% to 100%
* **Default**: **50%**
* **Effect**:
  * Higher values give more prominence to vector search results.
  * Lower values prioritize keyword search results.
  * Only applicable when Vector Search Mode is set to "Hybrid".

**Min Similarity Score**

* **Description**: Sets the minimum similarity threshold for vector search results.
* **Range**: 0% to 100%
* **Default**: **50%**
* **Effect**:
  * Higher values ensure only highly relevant vector results are included.
  * Lower values allow more diverse but potentially less relevant results.
  * Only applicable when Vector Search Mode is set to "Hybrid".

**Limit**

* **Description**: Sets the **maximum number of results** to return from a vector search.
* **Default**: **1000**
* **Type**: Integer
* **Note**: This setting applies to both Fallback and Hybrid modes.

## API Parameters

The following parameters can be used to override the vector search settings on a per-request basis:

**vectorMinSimi**

* **Description**: Overrides the minimum similarity score setting.
* **Range**: 0.0 to 1.0 (representing 0% to 100%)
* **Example**: `vectorMinSimi=0.7`

**hybridWeight**

* **Description**: Overrides the hybrid weight setting.
* **Range**: 0.0 to 1.0 (representing 0% to 100%)
* **Example**: `hybridWeight=0.8`

These parameters allow for dynamic adjustment of vector search behavior based on specific search contexts or user preferences.

> :test\_tube: **Beta Function**: if you need any further assistance please reach out to <service-desk@fact-finder.com>&#x20;


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.fact-finder.com/docs/modules/vector-search/hybrid-search.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.