PostgreSQL AI extension for vector search, embeddings, ML inference, RAG pipelines, MCP server, and agentic AI runtime. Runs in-process with SQL functions.
CREATE EXTENSION neurondb;
One platform. SQL-first.
Vector, ML, and RAG primitives that run where your data lives.
Vector search
HNSW and IVF indexing with multiple distance metrics. Compatible with pgvector.
SELECT * FROM vector_search('embeddings', q, 10);ML inference
In-database inference and classical ML algorithms as SQL functions.
SELECT neurondb.ml.predict('model', features);RAG pipeline
Chunking, retrieval, reranking, and answer assembly as database steps.
SELECT neurondb.rag.query('docs', q, top_k => 5);PostgreSQL-native
Runs in-process as an extension. No sidecars required.
CREATE EXTENSION neurondb;Agent runtime
Tool execution and workflows via NeuronAgent for database AI apps.
pip install neuronagentMCP tools
Model Context Protocol server for tool access from Claude Desktop and other clients.
npx @neurondb/neuronmcpNeuronDB Console
Manage your database and workflows from a unified interface
-- Vector similarity search
SELECT id, similarity, text
FROM vector_search('embeddings',
'[0.12, -0.45, 0.78, ...]'::vector(384),
5
);| id | similarity | text |
|---|---|---|
| 1042 | 0.9134 | kubernetes helm chart with prometheus grafana jaeger observability stack… |
| 991 | 0.9011 | hnsw index build parameters and ef_search tuning for optimal performance… |
| 807 | 0.8876 | hugging face dataset loader inserts embeddings into vector(384) columns… |
| 523 | 0.8754 | background workers for async embedding generation and indexing tasks… |
| 389 | 0.8621 | GPU acceleration for batch vector operations with CUDA support enabled… |
Trusted by teams building SQL-first AI — PostgreSQL-native vector search, ML inference, and RAG in one extension.
NeuronDB PostgreSQL AI Ecosystem
Complete AI stack for PostgreSQL
NeuronDB
PostgreSQL extension for AI
- •Vector search with HNSW/IVF/PQ/OPQ indexing
- •52+ ML algorithms in pure C
- •RAG pipeline with reranking
NeuronMCP
MCP server with 600+ tools
- •600+ MCP tools for database and AI operations
- •Claude Desktop integration
- •Vector search and ML inference tools
NeuronAgent
Agent runtime for workflows
- •Multi-step workflow execution
- •Tool call orchestration
- •State management and persistence
NeuronDesktop
Desktop app for management
- •Visual query builder
- •Agent workflow monitoring
- •Real-time performance metrics
NeuronDB Cloud
Managed PostgreSQL AI platform
- •Multi-cloud (AWS, GCP, Azure)
- •14+ microservices, Terraform & Helm
- •Stripe billing and metering
NeuronDB Hub
Agent builder and embed platform
- •No-code agent builder
- •Embeddable chat widget
- •Marketplace and MCP connectors
Inside NeuronDB
Engines, pipelines, and deployment tools
Vector engine
Index and search embeddings.
- •HNSW + IVF indexes
- •10+ distance metrics
- •GPU batch acceleration
- •pgvector compatible
ML engine
Inference and classical ML.
- •52+ algorithms
- •Model management
- •Feature store
- •Train in SQL
RAG pipeline
Retrieval + rerank.
- •Chunk + embed
- •Hybrid retrieval
- •Reranking
- •LLM integration
Background workers
Async tasks.
- •4 workers
- •Queue + scheduler
- •Metrics hooks
- •Async indexing
Kubernetes Helm
Cloud-native deployment.
- •StatefulSets
- •HPA + PDB
- •Prometheus + Grafana
- •Jaeger tracing
Hugging Face
Dataset loading.
- •Datasets API
- •sentence-transformers
- •Batch loading
- •Model caching
NeuronDB vs. Alternatives
Comparison of NeuronDB with other PostgreSQL AI and vector extensions
| Feature | NeuronDB | pgvector | pgvectorscale | pgai | PostgresML |
|---|---|---|---|---|---|
| Index types | HNSW, IVF, PQ, hybrid, multi-vector | HNSW + IVF | StreamingDiskANN | Uses pgvector | pgvector-based |
| ML Inference | ONNX (C++) | None | None | API calls | Python ML libs |
| Embedding Generation | In-database (ONNX) | External | External | External API | In-database (Transformers) |
| Hybrid Search | Native (Vector+FTS) | Manual | Manual | Manual | Manual |
| Reranking | Cross-encoder, LLM, ColBERT, MMR | None | None | None | None |
| ML Algorithms | 52 algorithms: RF, XGBoost, LightGBM, CatBoost, SVM, KNN, DT, NB, NN, K-means, DBSCAN, GMM, PCA, etc. | None | None | None | XGBoost, LightGBM, sklearn suite, Linear/Logistic |
| Background Workers | 4 workers: neuranq, neuranmon, neurandefrag, neuranllm | None | None | None | None |
| RAG Pipeline | Complete In-DB | None | None | Partial (API) | Partial (Python) |
| Quantization | FP16, INT8, Binary (2x-32x) | Binary only | Binary only | None | None |
| Implementation | Pure C | Pure C | Pure C | Rust + SQL | Python + C |
| GPU Support | CUDA + ROCm + Metal (native C/C++) | None | None | None | CUDA (via Python) |
| PostgreSQL Versions | 16, 17, 18 | 12-18 | 15-18 | 16-18 | 14-16 |
| Vector Types | 6 types: vector, vectorp, vecmap, vgraph, rtext, sparse_vector | 1 type: vector | 1 type: vector | Uses pgvector | Uses pgvector |
| Distance Metrics | 10+ metrics: L2, Cosine, Inner Product, Manhattan, Hamming, Jaccard, etc. | 3 metrics: L2, Cosine, Inner Product | 3 metrics: L2, Cosine, Inner Product | Uses pgvector | Uses pgvector |
| SQL Functions | 665+ functions | ~20 functions | ~30 functions | ~15 functions | ~50 functions |
| Benchmark coverage | RAGAS, MTEB, BEIR integrated | Manual setup required | Manual setup required | Manual setup required | Manual setup required |
| HNSW Index Build Performance | 606ms (50K, 128-dim L2, 10.1x), 583ms (50K, 128-dim Cosine, 8.8x), 146ms (10K, 768-dim L2, 27.1x), 1,208ms (100K, 128-dim L2, 13.0x) | 6,108ms (50K, 128-dim L2, baseline), 5,113ms (50K, 128-dim Cosine), 3,960ms (10K, 768-dim L2), 15,696ms (100K, 128-dim L2) | Varies by configuration | Uses pgvector | Uses pgvector |
| Performance (QPS) | 100K+ (with GPU) | 10K-50K | 50K-100K | Limited (API overhead) | 5K-20K (Python overhead) |
| Dependencies | Zero (pure C, optional ONNX) | Zero (pure C) | Zero (pure C) | Rust runtime | Python + ML libraries |
Use cases
Powering AI applications with PostgreSQL
Semantic search, RAG, recommendations, agents, and unified management — all from your database.
Semantic Search
Build semantic search engines with HNSW indexing and hybrid retrieval
RAG Applications
Create RAG pipelines with document processing, reranking, and LLM integration
Recommendation Systems
Build recommendation engines with collaborative filtering and ML algorithms
AI Agents
Deploy autonomous agents with long-term memory and tool execution via NeuronAgent
Claude Desktop Integration
Connect Claude Desktop to your database with 600+ MCP tools
Unified Management
Manage your entire NeuronDB ecosystem from a single dashboard
Managed NeuronDB (Cloud)
Run NeuronDB as a managed service on AWS, GCP, or Azure with multi-tenant auth and billing.
Embed AI Agents (Hub)
Build agents and embed a chat widget on any site with NeuronDB Hub.
FAQ
Frequently asked questions
Quick answers about NeuronDB and the PostgreSQL AI stack.
Get started
Build with SQL.
CREATE EXTENSION neurondb; then scale to production with Kubernetes and Helm. No external vector DB or ML service required.