Percona Search for MongoDB documentation¶
Technical Preview
Percona Search for MongoDB 1.70.3-1 is available as a technical preview.
We recommend that early adopters use this release for testing purposes only and not in production environments.
Percona Search for MongoDB adds full-text and vector search to self-managed Percona Server for MongoDB deployments. Use it to build keyword, semantic, and hybrid search, along with recommendation and retrieval-augmented generation (RAG) applications.
Search runs in a separate process called mongot. This process maintains Apache Lucene-based search indexes and executes the queries that you submit through the $search, $searchMeta, and $vectorSearch aggregation stages.
What’s new in Percona Search for MongoDB 1.70.3-1
Get started¶
Check the requirements, start the mongot service, connect it to Percona Server for MongoDB, and verify that the two components can communicate.
Understand the architecture¶
See how mongod, mongot, and your search indexes work together, from a single query’s path through the system to how a sharded cluster splits the work.
Create and manage search indexes¶
Create, check the status of, update, and delete full-text and vector search indexes as your data and query needs change.
Build search experiences¶
Query by keyword, by semantic meaning, or combine both in a single pipeline. Percona Search for MongoDB currently works with embeddings you generate yourself or automatically generated, giving your application full control over the embedding model and workflow.