FAISS Vector DB Search

Overview

FAISS Vector Search provides high-performance semantic similarity search capabilities for FAISS vector databases. This service leverages FAISS's optimized algorithms to find the most semantically similar segments to user queries, with configurable precision and performance parameters.

Service Information

Property	Value
Service Name	Search FAISS vector DB
Status	Enabled
Compatible Nodes	Search knowledge DB

When to Use This Service

FAISS Vector Search is ideal for:

Local semantic search applications with fast response requirements
Development and testing of RAG systems with controlled infrastructure
High-performance search where millisecond response times are critical
Cost-sensitive applications avoiding cloud service fees for search operations
Offline search capabilities without external service dependencies

Parameters

Parameters automatically adapt based on the FAISS index type used when creating the database:

Inverted Index Parameters

Parameter	Type	Default	Description
Number of results	Number	5	Maximum number of search results to return
Number of probes	Number	5	Number of index cells to probe during search (higher = more accurate, slower)

HNSW Index Parameters

Parameter	Type	Default	Description
Number of results	Number	10	Maximum number of search results to return
Search effort	Number	200	Search quality vs. speed trade-off (higher = more accurate, slower)

How It Works

Query Processing: Converts user prompt to embedding vector using the same model used for database creation
Index Search: Searches FAISS index using configured parameters and similarity metric