Lantern offers an open‑source Postgres extension suite that adds scalable vector search, BM25 text ranking, hybrid search, and on‑database embedding generation and LLM inference to a standard PostgreSQL instance. By using pgvector‑compatible HNSW indexes and optional serverless indexing, developers can handle millions to billions of vectors with low latency while keeping all data and queries within a single database, and they can choose between self‑hosting or a managed Lantern Cloud service.
Funding
$500K raised to dateRaised to date based on public sources. This may differ from the amount the company actually raised and is based only on what is publicly available on the internet.
Founders
Product
Problem
Developers building AI applications often need to add vector search, embedding generation, and LLM inference to their data pipelines, which typically requires managing separate specialized databases, search engines, and external services. This adds operational complexity, latency, and scaling challenges, especially when handling millions of vectors.
Solution
Lantern provides an open-source Postgres extension suite that brings vector search, BM25 text ranking, hybrid search, and on‑database embedding generation into a single Postgres instance. By leveraging pgvector‑compatible HNSW indexes and serverless external indexing, Lantern can scale to billions of vectors while keeping the primary database responsive. The platform also offers managed cloud instances (Lantern Cloud) and support for self‑hosting, allowing users to create, tune, and query vector indexes with standard SQL. Integrated LLM functions let developers generate embeddings or run completions directly in the database, eliminating the need for separate inference services. Pricing is tiered by compute and storage, with a free tier for experimentation and paid plans for production workloads.
Target Audience
Primary customers are software engineers and data teams building AI‑enhanced applications—such as recommendation engines, code search, and enterprise knowledge bases—that require high‑performance vector search and embedding generation within their existing Postgres databases.
Features
- pgvector‑compatible HNSW index creation that scales to billions of vectors, with optional serverless indexing to offload heavy indexing work
- Built‑in BM25 ranking for high‑quality sparse text search and hybrid search combining dense vectors and BM25 via RRF or custom reranking
- SQL‑level embedding generation and LLM calls supporting multiple open‑source and commercial models (e.g., OpenAI, Cohere, BAAI)
- Automatic index hyperparameter optimization and support for binary, scalar, and product‑compressed vectors
- Seamless integration with existing Postgres schemas; add vector columns and queries using standard SQL operators
- Managed Lantern Cloud service with multi‑tenant free tier and scalable production instances (standard‑1 to standard‑16)
- CLI and extension packages (Lantern HNSW, Lantern Extras) for self‑hosting or deployment on Ubicloud