Vespa (7,037 Stars) 2026: The Big Data Serving Engine for Vector Search, Ranking and Recommendations

Vespa (7,037 stars) is Yahoo's open-source serving engine for vector search, real-time ranking and recommendations at scale. Here is how to deploy it and run hybrid search.

💡 What You Will Learn

Vespa (7,037 stars) is Yahoo's open-source serving engine for vector search, real-time ranking and recommendations at scale. Here is how to deploy it and run hybrid search.

📜 Table of Contents

• The short answer
• What Vespa does uniquely
• Deploy locally
• A minimal hybrid search schema
• Practical tips
• FAQ

The short answer

vespa-engine/vespa (7,037 stars, Java) is an open-source big data serving engine - the technology behind Yahoo Search's ranking. It handles vector search, BM25, and machine-learned ranking in one system, with real-time writes and sub-second query latencies at massive scale.

What Vespa does uniquely

• Hybrid search: combine dense vectors and BM25 in one query
• Real-time indexing: documents are searchable immediately after write
• ML ranking: rank with ONNX models at query time
• Serving: designed for production traffic, not just offline experimentation

Deploy locally

docker run --detach --name vespa --hostname vespa-container \n  --publish 8080:8080 --publish 19071:19071 \n  vespaengine/vespa

# Deploy a sample app
vespa deploy sample-apps/vector-search
vespa query 'select * from doc where userQuery()' 'input.query(q)=embed(foo)'

Vespa CLI (brew install vespa-cli) drives deploy and query workflows.

A minimal hybrid search schema

schema doc:
  document doc:
    field text type string { indexing: index }
    field embedding type tensor<float>(x[384]) {
      indexing: attribute | index
    }
  rank-profile hybrid inherits default {
    first-phase {
      expression: bm25(text) + cos(distance(field, embedding))
    }
  }

Practical tips

• Use Vespa when you need vector + keyword + learned ranking together at scale.
• Start with the sample apps; production setups use config server + content nodes.
• Monitor with the built-in metrics and logs in the admin console.

FAQ

Who uses it? Yahoo, Spotify, and many others run it in production for search and recommendations.

Is it a vector database? It is a full serving engine: vector DB + search + ranking in one.

Is it free? Yes - Apache-2.0 open source.

❓ FAQ

Who uses it?

Yahoo, Spotify, and many others run it in production for search and recommendations.

Is it a vector database?

It is a full serving engine: vector DB + search + ranking in one.