Lantern Cloud

Open-source Postgres vector database and toolkit for scalable AI applications.

Freemium

Description

Lantern Cloud is an open-source vector database built on Postgres, enabling developers to build and scale production-ready AI applications efficiently. It combines advanced vector and hybrid search, allowing seamless handling of both dense and sparse data within a familiar Postgres environment.

With support for serverless indexing, BM25 text search, and native integration of embedding generation and LLM models, Lantern Cloud streamlines AI workflows by letting users generate and utilize vectors directly via SQL. Choose between self-hosting for full control or managed cloud for increased scalability, performance, and reliability.

Key Features

Vector Search with pgvector: Search over dense and sparse vectors efficiently within Postgres.
BM25 Text Search: Enhance search relevance with BM25 ranking, used in ElasticSearch.
Hybrid Search: Combine vector and text search using advanced reranking algorithms.
Serverless Indexing: Offload vector index creation for scalability without performance loss.
In-Database Embedding and LLM Generation: Generate embeddings or run LLM models via simple SQL commands.
Support for Multiple Models: Integrates with OpenAI, Cohere, Jina AI, and more than 20 embedding and LLMs.
Automatic Vector and LLM Columns: Easily create vector or LLM columns from existing data.
Easy Integration: Use pure SQL or popular ORMs like Sequelize, Knex, and Django.
Infinite Scalability: Flexible compute resources and managed or open-source deployment.
Managed or Self-Hosted Options: Choice between fully-managed cloud or self-hosting for control.

Use Cases

Building scalable retrieval-augmented generation (RAG) applications
Semantic and full text search over large datasets
In-database embedding and LLM workflow automation
Scaling codebase or repository search across millions of files
Real-time AI-powered data analytics
Knowledge base and enterprise search engine development
Recommendation system backend
Search over product catalogs and business data

Frequently Asked Questions

Is Lantern Cloud open-source?

Yes, Lantern Cloud is open-source, allowing for self-hosting or use as a managed service.

What is included in the free tier?

The free tier includes a multi-tenant setup with vector search and indexing, vector and LLM generation, automatic upgrades, query insights, and community support.

Which embedding and LLM models are supported?

Lantern Cloud supports over 20 models including OpenAI, Cohere, Jina AI, BAAI, and open-source models.

How does serverless indexing work?

Serverless indexing allows vector indexes to be created on a separate machine from your main database, enabling scalability without affecting database performance.

Can I integrate Lantern Cloud with my ORM or use SQL directly?

Yes, Lantern Cloud supports SQL queries and integrates with popular ORMs such as Sequelize, Knex, and Django.