Vespa

Vespa lets you build search and AI applications by combining search, ranking, machine learning inference, and real-time updates in one platform. It can query, organize, and run inference over vector, tensor, text, and structured data, and is described as handling thousands of queries per second with latency under 100ms over billions of constantly changing data items. It distributes large amounts of data across multiple nodes and evaluates them in parallel, with high availability and high performance. You can deploy to Vespa Cloud or install and run it yourself.

• Vector, text, and structured search
• Distributed machine learning ranking (TensorFlow, ONNX, XGBoost, and LightGBM models)
• RAG using hybrid search, relevance models, and multi-vector
• Streaming search mode for personal data
• Auto-scaling and continuous deployment and upgrades
• Fully managed cloud with strong security
• Deploy in your own cloud (Enclave: AWS, Azure, GCP)

• One of 12 AI tools in the Vector & Search field.
• Connected to 2 other tools — #7 in its field.