NeuronDB FAQs

Question 1

What AI features does NeuronDB add to PostgreSQL?

Accepted Answer

NeuronDB extends PostgreSQL with in-database vector search using HNSW and IVF indexing, a comprehensive suite of 50+ machine learning algorithms (classification, regression, clustering), and capabilities for generating text, image, and multimodal embeddings.

Question 2

How does NeuronAgent enhance NeuronDB's capabilities?

Accepted Answer

NeuronAgent is a component of the NeuronDB ecosystem that provides an agent runtime system with REST API and WebSocket endpoints. It enables building autonomous applications with long-term memory, tool execution, and streaming responses, all leveraging the AI power of NeuronDB.

Question 3

Does NeuronDB support GPU acceleration for AI operations?

Accepted Answer

Yes, NeuronDB is engineered for high performance and supports GPU acceleration across multiple platforms including CUDA (NVIDIA), ROCm (AMD), and Metal (Apple) for its AI operations, significantly speeding up vector searches, ML model training, and inference.

Question 4

Can NeuronDB be used for Retrieval Augmented Generation (RAG) applications?

Accepted Answer

Absolutely. NeuronDB includes a robust RAG pipeline with LLM integration, making it ideal for building applications that require efficient document retrieval, context generation, and hybrid search combining vector and full-text search capabilities directly within PostgreSQL.

Question 5

What is NeuronDB?

Accepted Answer

NeuronDB is a powerful PostgreSQL extension that integrates advanced AI capabilities directly into your database. It provides vector search, over 50 machine learning algorithms, embedding generation, and an agent runtime for building intelligent, autonomous applications.

NeuronDB

NeuronDB

主要功能

使用案例

主要功能

使用案例