Complete documentation for the NeuronDB ecosystem. Build AI applications in PostgreSQL with vector search, ML inference, and RAG pipelines.
$ docker run -d --name neurondb \
-e POSTGRES_PASSWORD=postgres \
neurondb/neurondb:latestSelect a Product
NeuronDB
PostgreSQL AI Extension
PostgreSQL AI extension with vector search, ML inference, RAG pipeline, and GPU acceleration. Build AI applications directly in PostgreSQL.
NeuronAgent
AI Agent Runtime for PostgreSQL
REST API and WebSocket agent runtime system with long-term memory, tool execution, and streaming responses.
NeuronMCP
Model Context Protocol Server for NeuronDB
Model Context Protocol server with 600+ tools (100+ PostgreSQL admin + 100+ vector operations + ML, RAG, enterprise, monitoring, multimodal, workflow, dataset loading) enabling MCP-compatible clients (like Claude Desktop) to access NeuronDB vector operations, ML training, RAG pipeline, and PostgreSQL administration.
NeuronDesktop
Unified Web Interface for NeuronDB Ecosystem
Unified web interface providing a single dashboard for managing and interacting with MCP servers, NeuronDB, and NeuronAgent. Real-time WebSocket communication, secure authentication, and comprehensive monitoring.
NeuronDB Cloud
Managed NeuronDB on Kubernetes
SaaS platform on Kubernetes delivering three integrated managed services per tenant: NeuronDB (vector search, ML, RAG), NeuronAgent (agent runtime), and NeuronMCP (600+ tools). Multi-cloud (AWS, GCP, Azure), hybrid tenancy, Terraform and Helm, Stripe billing, RBAC, audit logging, backup/PITR, and CLI for tenants and usage.
NeuronDB Hub
Agent Builder and Embed Platform
Hosted agent builder, control plane for NeuronAgent, secure gateway for runtime, and one-click embed platform. Multi-tenant SaaS: create agents, manage knowledge and MCP connectors, run workflows and memory, embed a chat widget on any site. TypeScript and Python SDKs; Stripe billing; optional NeuronDB Cloud integration.