Vespa python API

Vespa is the scalable open-sourced serving engine to store, compute and rank big data at user serving time. pyvespa provides a python API to Vespa.

We aim for complete feature parity with Vespa, and estimate that we cover > 95% of Vespa features, with all most commonly used features supported.

If you find a Vespa feature that you are not able to express/use with pyvespa, please open an issue.

Quick start

To get a sense of the most basic functionality, check out the Hybrid Search Quick start:

• Hybrid search quick start - Docker
• Hybrid search quick start - Vespa Cloud

Overview of pyvespa features

Info

There are two main interfaces to Vespa:

1. Control-plane API: Used to deploy and manage Vespa applications.
   • VespaCloud: Control-plane interface to Vespa Cloud.
   • VespaDocker: Control-plane iterface to local Vespa instance (docker/podman).
2. Data-plane API: Used to feed and query data in Vespa applications.
   • Vespa

Note that VespaCloud and Vespa require two separate authentication methods.

Refer to the Authenticating to Vespa Cloud for details.

• Create and deploy application packages, including schemas, rank profiles, services.xml, query profiles etc.
• Feed and retrieve documents to/from Vespa, using /document/v1/ API.
• Query Vespa applications, using /search/ API.
• Build complex queries using the QueryBuilder API.
• Collect training data for ML using VespaFeatureCollector.
• Evaluate Vespa applications using VespaEvaluator/VespaMatchEvaluator.

Requirements

Install pyvespa:

We recommend using uv to manage your python environments:

uv add pyvespa

or using pip:

pip install pyvespa

Check out the examples

Check out our wide variety of Examples that demonstrate how to use the Vespa Python API to serve various use cases.