Vespa Review, Pricing & Features

Vespa is an open-source engine for search, recommendation, and real-time ML inference at scale. Learn its history, features, and Vespa Cloud pricing.

Category
Databases
Pricing
open-source, from Free (open-source, self-hosted); Vespa Cloud is usage-based with a free trial
Verified
Not yet
Last updated
July 19, 2026
Founded
2023
Headquarters
Trondheim, Norway
APIOpen SourceAISelf-Hosted

What Is Vespa

Vespa is an open-source, distributed serving engine for search, recommendation, and real-time machine learning inference at large scale. Its technology originated at Yahoo in 2005, built out of the AllTheWeb search team following Yahoo's acquisition of Overture, and was used internally across Yahoo's products for over a decade before being open sourced in 2017.

External companies including Spotify and Wix began adopting Vespa around 2021. On October 4, 2023, Yahoo spun Vespa out as a fully independent company, appointing longtime Yahoo big-data architect Jon Bratseth as CEO, while retaining an equity stake and board seat.

The independent company is headquartered in Trondheim, Norway, with roughly 60 to 70 employees, and raised a 31 million dollar Series A from Blossom Capital in November 2023, with Yahoo and investor J12 also participating.

Key Features

Vespa unifies vector search, tensor operations, full-text search, and structured data queries in a single engine, which is increasingly relevant for AI applications that combine semantic and keyword search, such as retrieval-augmented generation.

The engine is designed to scale to billions of constantly changing documents while serving thousands of queries per second at sub-100-millisecond latency, using a distributed, machine-learned ranking pipeline.

Vespa supports continuous deployment and automated scaling for production workloads, and Vespa Cloud extends this with fully managed hosting, security, and operations on top of the open-source engine.

Pricing

The core Vespa engine is free and open source, and teams can self-host it on their own infrastructure at no licensing cost.

Vespa Cloud, the managed hosting option, uses usage-based pricing rather than a flat published rate, and offers a free trial for teams evaluating the managed service before committing to a paid plan.

Key Features

Pros & Cons

Pros

  • Proven at massive scale from over a decade of internal Yahoo production use
  • Combines vector search, text search, and ML ranking in one engine without separate systems
  • Free, open-source core with no vendor lock-in
  • Strong performance at high query volume and low latency
  • Backed by an experienced team and disclosed venture funding for continued development

Cons

  • Steeper learning curve than simpler, purpose-built vector databases
  • Smaller community and ecosystem than Elasticsearch or newer vector-only databases
  • Vespa Cloud pricing is not published and requires contacting sales
  • Self-hosting at scale requires meaningful operational expertise

Pricing

Frequently Asked Questions

What is Vespa.ai used for?

Vespa is used for large-scale search, recommendation, and AI serving applications.

Is Vespa free?

Yes, the core engine is open source and free to self-host; Vespa Cloud is a paid, usage-based managed option.

Who created Vespa?

Vespa was originally built at Yahoo starting in 2005 and open sourced in 2017.

When did Vespa become an independent company?

Yahoo spun Vespa out as an independent company on October 4, 2023.

Who leads Vespa.ai today?

CEO Jon Bratseth, a former Yahoo big-data architect and core Vespa contributor, leads the company.

Does Vespa support vector search for AI applications?

Yes, alongside text search and structured data queries in the same engine.

Is Vespa related to the Vespa scooter brand?

No, it is entirely unrelated; the shared name is coincidental.

Which companies use Vespa?

Spotify and Wix are among the known external companies that use Vespa.

Related Tools