Vespa

Vespa is an open-source serving engine for text, vectors, tensors, and structured data, plus a managed Cloud. Rechecked 2026-08-18: github.com/vespa-engine/vespa had 7,052 stars, license Apache-2.0, homepage vespa.ai. vespa.ai still sells search, RAG, recommendation, and online ranking together. docs.vespa.ai/en/cloud and cloud.vespa.ai describe Vespa Cloud. cloud.vespa.ai/en/pricing has a unit table, not a $20 seat. vespa.ai/free-trial and the README still say you can run apps in Cloud for free during a trial. Drop leftover "full-featured search+vector service" copy that never printed OSS stars or Cloud units.

Compare Qdrant if you wanted a dedicated vector engine with Cloud Free / usage-based Standard, Pinecone if you wanted Free / $20 / $50 min / $500 min, and Milvus if you wanted another OSS vector database.

Key Features

• OSS plus Cloud: Self-host Apache-2.0 Vespa, or deploy on Vespa Cloud / Vespa on AWS. Do not treat the marketing homepage as a dollar quote.
• Hybrid retrieval, not vectors only: First-party pages still combine keyword, vector, tensor, and structured search with online ML ranking. That is the product, not a leftover algorithm zoo.
• Cloud prices are units: cloud.vespa.ai/en/pricing lists vCPU hour, memory GB hour, disk GB hour, GPU hour, volume GB month, and network GB. AWS and GCP rows differ. Recheck that page before you invent a seat.
• Free trial is not a forever Free node: vespa.ai/free-trial and the README advertise a Cloud trial. We did not treat that as Qdrant-style free-forever.

Use Cases

• People who will self-host Vespa and pay their own machines.
• Teams that need ranking + hybrid search, not only ANN.
• Teams that will open the Cloud calculator and accept unit billing.

Limitation: vespa.ai/pricing is a marketing page without a printed seat. We did not invent a dollar-per-month plan.

Pricing

GitHub, vespa.ai, and cloud.vespa.ai/en/pricing on 2026-08-18.

If a leftover blog still calls Vespa an unpriced all-in-one toolkit, treat this table as the source.

Getting Started

1. Open github.com/vespa-engine/vespa or docs.vespa.ai.
2. Recheck cloud.vespa.ai/en/pricing before you assume a $20 Builder row.
3. Recheck Qdrant if you wanted a printed Cloud Free node.
4. Do not paste homepage "below 100 milliseconds" as a contract.

Frequently Asked Questions

Is the engine free?

Yes. Apache-2.0. Cloud is a separate product with a trial and unit billing.

Is there a public dollar table?

There is a Cloud unit table. There is no public $20 / $50 / $500 seat matrix like Pinecone.

Same as Qdrant?

No. Qdrant is a vector engine plus Cloud Free / usage-based Standard. Vespa is a serving platform that also does vectors.

Alternatives

• Qdrant: OSS vector engine plus Cloud Free / usage-based Standard.
• Pinecone: Managed vector DB with public dollar rows.
• Milvus: Another open vector database; managed Cloud is Zilliz.

Tips

1. Quote 7,052 stars and Apache-2.0 from the GitHub check.
2. Split Vespa OSS from Vespa Cloud units.
3. Recheck the Cloud calculator before you invent a monthly seat.

Conclusion

Vespa is an Apache-2.0 search and serving engine with 7,052 stars; Cloud bills units after a trial, not an unpriced toolkit. Start at github.com/vespa-engine/vespa or cloud.vespa.ai/en/pricing, then decide whether Qdrant already covers the vector Cloud you need.