vespa-engine/vespa

Health v2 · maintenance only

Adoption not part of the score

7069 stars · 735 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

Vespa is an open-source, distributed AI search platform and serving engine that combines full-text search, vector/tensor search, and machine-learned ranking with real-time inference over large, continuously changing datasets. It scales to billions of documents and hundreds of thousands of queries per second with sub-100ms latencies, and can be self-hosted or run as a managed cloud service.

Use cases

• build a hybrid search engine combining BM25 text search with vector similarity
• serve RAG applications with retrieval and embedding inference at scale
• run a vector database for nearest-neighbor search over billions of embeddings
• build personalized recommendation and ranking systems with ML model inference at query time
• deploy low-latency search over constantly updating large corpora
• evaluate and rank results with ONNX, TensorFlow, XGBoost, or LightGBM models at serving time

When to choose

• you need production-scale search, recommendation, or RAG with sub-100ms latency over billions of documents
• you want unified full-text, vector, and structured data search with integrated ML ranking in one platform
• you need real-time indexing and querying of continuously changing data
• you want to self-host a battle-tested engine proven on large internet services like Perplexity

When to avoid

• you need a simple embedded search library for a small application without distributed serving
• your team cannot operate a complex multi-node Java-based platform and you don't want a managed cloud
• you only need lightweight keyword search where Elasticsearch or a simpler tool suffices
• you want a pure vector store without text search or ranking features

Facets

service · maturity stable

search-engine vector-database machine-learning rag llm-inference streaming databases large-language-models big-data machine-learning analytics cloud self-hosted jvm python hybrid-search tensor-computation ranking recommendation distributed-serving approximate-nearest-neighbor bm25 real-time-indexing pyvespa search retrieval-augmented-generation linux macos docker kubernetes web-server

Member repositories

For agents

markdown · JSON · MCP: product_card(name="vespa-engine/vespa")