Open Source Vector Database

Vector data querying with SQL

Hyper-fast. Queries in milliseconds.

        
SELECT text, _score
FROM word_embeddings
WHERE knn_match(embedding,[0.3, 0.6, 0.0, 0.9], 2)
ORDER BY _score DESC;

        
|------------------------|--------|
|         text           | _score |
|------------------------|--------|
|Discovering galaxies    |0.917431|
|Discovering moon        |0.909090|
|Exploring the cosmos    |0.909090|
|Sending the mission     |0.270270|
|------------------------|--------|

        
SELECT text, _score
FROM word_embeddings
WHERE knn_match(embedding, (SELECT embedding FROM word_embeddings WHERE text ='Discovering galaxies'), 2)
ORDER BY _score DESC

        
|------------------------|--------|
|         text           | _score |
|------------------------|--------|
|Discovering galaxies    |1       |
|Discovering moon        |0.952381|
|Exploring the cosmos    |0.840336|
|Sending the mission     |0.250626|
|------------------------|--------|

Streamlined data management

Eliminate the need to manage multiple systems. CrateDB seamlessly integrates your data, keeping your (meta-)data and vector representations aligned without the complexity of data synchronization processes. Not only does it offer powerful vector search capabilities, but it also seamlessly integrates with time series, geospatial, JSON, full-text search, and other data types.

cr-quote-image

Data enriched with semantics

Seamlessly add vector data types to any row in the database, providing context aligned with your (meta-)data and enhancing explainability.

cr-quote-image

Advanced search capabilities

Combine vector, full-text, and keyword searches for improved semantic similarity and keyword matching, enhancing search precision and relevance.

cr-quote-image

Enhanced AI model integration

Leverage CrateDB's native support for complex data analytics to speed up integration with AI models, optimizing your AI projects.

cr-quote-image

Improved scalability

Handle vector data with confidence, eliminating the need for separate vector databases and enabling smoother scaling as your data grows.

cr-quote-image

Faster development & lower maintenance

Save time in development by avoiding the integration and learning curve associated with external solutions for vector data storage and retrieval.

cr-quote-image

Learn more about the product

Keynote - The transformative effects of real-time AI

Keynote - The transformative effects of real-time AI

In this keynote at the AI & Big Data Expo Europe 2023, CrateDB's VP Product shares his vision for the future with multi-model SQL databases and Large Language Models.

Dev Talk - How to use private data in generative AI

Dev Talk - How to use private data in generative AI

This talk at Fosdem 2024 focuses on the combination of CrateDB and LangChain: it helps get started with using private data as context for large language models through LangChain, incorporating the concept of Retrieval Augmented Generation (RAG).

5 essential things you need to know about vector databases

Infographic-Vector-Database-Essentials

5 essential things you need to know about vector databases

This infographic gives you some basic understanding of vector databases, from what you should look for when choosing one to combining vector data with other data types.

Demo – Harnessing CrateDB’s multi-model capabilities for AI-powered applications

Demo – Harnessing CrateDB’s multi-model capabilities for AI-powered applications

In this video, we explore the integration of CrateDB and PyCaret to detect anomalies in machine data, crucial for identifying potential failures or inefficiencies in technological systems. CrateDB's capability for handling large-scale data with ease pairs seamlessly with PyCaret's low-code approach to machine learning, offering a streamlined path to uncovering insights within vast datasets.

CrateDB stands as a vector store database with key features that elevate its capabilities: vector storage and similarity search.

Vector storage empowers users to efficiently store embeddings produced by their preferred machine learning models, creating a streamlined method for managing and accessing vectorized data.
Similarity search enables users to effortlessly discover similarities within datasets represented as vectors, fostering advanced data exploration and in-depth analysis.

By offering these vector database capabilities within a single, scalable product, CrateDB streamlines data management, cutting down both development time and total cost of ownership.

E-commerce recommendations

Elevate e-commerce experiences by enabling semantic vector and full-text searches to deliver more relevant product recommendations.

cr-quote-image

Chatbots & customer support

Enhance customer interactions by understanding questions with precision. Contextualize conversations, providing better service with improved understanding of user inquiries, regardless of the terms they use.

cr-quote-image

Anomaly & fraud detection

Detect outliers effortlessly in various sectors, from finance to IoT and cybersecurity by normalizing and encoding user or machine behavior into vectors.

cr-quote-image

Multimodal search

Extend your search capabilities beyond text to include images and videos.

cr-quote-image

Generative AI

Store embeddings, provide additional context in prompts and act as conversational memory for LLM-based applications. Use vector search functionality for retrieval augmented generation (RAG), which enables LLMs to understand specific data.

cr-quote-image

Vector data allows users to capture the complex details of points, lines, and polygons, unveiling a new dimension in data analysis, mapping, and spatial decision-making. Vector data can be stored in different file formats: Shapefile (.shp), GeoJSON (.geojson), KML (Keyhole Markup Language), and GML (Geography Markup Language). CrateDB leverages the power of vector data with a highly scalable database that can be queried using SQL, simplifying data management and reducing development time and overall costs.

Vectors are numerical representations used to quantify and compare features or characteristics of data items, such as text, images, or sounds, in a high-dimensional space. For example, a vector can look like this: -0.32643065, -0.12308089, -0.2873811, representing a point in a multi-dimensional space. In CrateDB, vectors are stored as one-dimensional arrays of float values using the float_vector data type, allowing for efficient storage and querying of dense vector data.

A vector database is designed to store and manage high-dimensional data, grouping vectors based on their similarities. These databases use advanced indexing and search algorithms to find the most similar vectors to a given query quickly. Examples of vector databases include CrateDB, Pinecone, Zilliz, and Weaviate. CrateDB excels as a vector store database with features like vector storage and similarity search. If you want to learn more, read this blog post on how to choose the best database for vector data.

Vector data is used in various applications, including e-commerce recommendations, chatbots and customer support, anomaly and fraud detection, multimodal searches, and generative AI. One of the key applications is similarity search, where algorithms like k-nearest neighbors (KNN) identify the most similar data points to a given query vector. This capability is crucial for recommendation systems, image retrieval, and anomaly detection. CrateDB enhances these applications by integrating vector storage and similarity search within a scalable database solution. Watch CrateDB’s keynote to learn more about vectors for real-time AI >

Typical distance metrics for comparing vectors include Euclidean distance, Cosine similarity, and Manhattan distance. In AI and ML, they are used for similarity search.