Search overview¶

Percona Search for MongoDB adds full-text search and vector search to Percona Server for MongoDB (PSMDB) 8.3 and later. It runs mongot, a dedicated search service, alongside mongod to build search indexes on your collections and process search queries.

Technical Preview

Percona Search for MongoDB is available as a technical preview. We recommend that early adopters use it for testing and evaluation only, and not in production environments.

You can create search indexes on your collections and use aggregation pipeline stages such as $search, $searchMeta, and $vectorSearchto retrieve relevant results.

What is mongot?¶

mongot is a companion process that builds and maintains search indexes for your MongoDB collections. While mongod stores and manages your application data, mongot creates optimized search indexes and processes search queries.

The two services communicate internally during query execution:

mongod stores documents and handles database operations.
mongot maintains search indexes.
Search queries are processed by mongot, while mongod retrieves the matching documents and returns the results to the client.

This architecture separates search workloads from core database operations while keeping the indexed data synchronized with the database.

Search types¶

Percona Server for MongoDB supports the following search types. Choose the type that matches your query patterns:

Search type	Query stages	Index type	Use for
Full-text search	`$search`, `$searchMeta`	`search`	Relevance-ranked text queries, autocomplete, faceting, and highlighting
Vector search	`$vectorSearch`	`vectorSearch`	Semantic similarity queries using machine learning embeddings

Full-text search¶

Full-text Search lets you search text stored in one or more fields using relevance-based ranking. It supports capabilities such as:

Keyword and phrase searches
Boolean operators
Fuzzy matching
Wildcard searches
Field-specific searches
Relevance scoring

Use cases include:

Product catalog search
Documentation search
Blog and article search
Customer support knowledge bases

Vector search¶

Vector search finds documents based on the meaning behind your data, rather than exact keyword matches. It compares the vector embedding of your query against the embeddings stored in your collection, and returns the documents whose embeddings are closest in meaning. It supports capabilities such as:

Approximate nearest neighbor (ANN) search, which uses an HNSW index to find close matches quickly across large datasets
Exact nearest neighbor (ENN) search, which compares every indexed vector for guaranteed accuracy on smaller datasets
Pre-filtering, so you can narrow the documents considered before the vector comparison runs
Manual embeddings, so you can generate vector embeddings with an embedding model of your choice and store them alongside your documents