The most powerful vector database for building AI applications.

Open-source PostgreSQL database extension for vector data and vector search operations

PostgreSQL built for AI applications

Unlike standalone vector engines, Lantern enables seamless combination of relational data and vector data for applications. Tap into the power of embedding models and large language models to easily build data-driven applications.

The best performance and 100% open-source

Lantern benchmarks outperform pgvector, the only other PostgreSQL extension on the market. Our commitment to open-source means Lantern is free for all. Benefit from the ever-growing enhancements and features that developers passionately contribute to.

Scale your applications without compromises

Extensions like pgvector utilize IVFflat, an algorithm that leads to clunky index management requirements and poor performance at scale. We built a vector database using a better algorithm, HNSW, to enable better throughput, latency, recall, and scalability.

AI applications forevery enterprise,for every industry.

Legal

Document Recommendation

Index legal documents like contracts, briefs, filings etc. by passing them through an LLM to generate vector embeddings. Vectorize legal document drafts to find similar documents that may lead to better final products.

Enterprise

Data augmented Q&A

Ingest your private company data, vectorize and store your data in Lantern. Build applications that allow Large Language Models like OpenAI to reference the data and generate responses

Finance

Fraud Detection

Embed transactional data, like past legitimate and fraudulent transactions such as purchase history, web/app activity, geolocation, etc. Use vector search to determine if new transactions are fraudulent.

Coming soon to

Self-hosted enterprise solution

Integrations with LangChain and LlamaIndex

Cloud hosted Lantern database

Industry specific templates for building applications

Benchmarking key performance metrics against prominent vector databases

Frequently Asked Questions

    Semantic search is the concept of searching on the meaning of data rather than the data itself.
    A vector is a numeric representation of your data that can be searched over using advanced machine learning algorithms.
    Pinecone requires you to adopt an entirely new database to build vector applications. Lantern extends your existing database, so that a separate database is not needed.
    Lantern supports embeddings from any provider
    Lantern will soon be supported as a Vector Store in both LangChain and LlamaIndex, two popular frameworks for building services that utilize Large Language Models.