Percona Launches Search for MongoDB
The source-available tool adds full-text, vector and hybrid search to MongoDB databases on an organisation's own infrastructure.
Updated on Nov. 4, 2026 in Artificial Intelligence
Percona has released Percona Search for MongoDB, a source-available search tier supporting full-text, vector and hybrid search. The tool runs on the organisation's own infrastructure and generates embeddings locally.
Why it matters
Local embedding generation can avoid per-token costs from external services that rise as data and query volumes grow. It may also help teams with strict data-handling policies take search projects into production.
Percona Search for MongoDB is built on mongot, MongoDB's vector, full-text and hybrid search engine, and integrates with Percona Server for MongoDB (PSMDB) and the Percona Operator. Performance metrics and benchmark results were not reported in the source material.
The players
Percona
Percona is an open-source database software and services company best known for its MySQL, MongoDB and PostgreSQL distributions and its Percona Operators for Kubernetes.
The details
The tool is free to use with no paid edition, deployment limits or license-gated features, though an optional add-on to a Percona Support contract covers the search tier. The Percona Operator deploys and scales the search tier independently of the database tier, and Percona plans future support for Percona Backup for MongoDB and Percona Monitoring and Management.
Timeline
• November 4, 2026: Percona Search for MongoDB released and reported as generally available.
• October 2026: Percona announced development of Valkey-proxy, an open-source proxy connecting applications to Valkey clusters.
Under the Hood
Percona's release builds directly on MongoDB's mongot engine, which powers the vector, full-text and hybrid search capabilities now offered in Percona Search. It positions self-hosted search as an alternative to cloud-managed search services that charge per use.
Developers running MongoDB can add vector and full-text search without sending data to external services, which may reduce per-token embedding costs. Teams can deploy the search tier through the Percona Operator, scaling it separately from the database.
The takeaway
Self-hosted search tiers give teams more control over data and costs, but they also shift the work of running embeddings and scaling search onto their own infrastructure.
Further reading
For more on the vector and embedding tools shaping database search, see Artificial Intelligence coverage.
