An open-source database for vector embeddings

A faster, cheaper and easier way to add semantic search into your app.

Twice the storage of Pinecone, at half the cost.

Star on GitHub 88

Why citrus

A better vector db

We're building a better vector database that gives you more control over your data, improved metrics and observability and overall, less headaches.

Open source
Citrus is completely open-source. Self-host it within your network or contribute to citrus directly!
Scalable
Citrus managed infrastructure scales with your data. Zero headaches when it comes to maintaining and updating the infra.
Cost efficient
High performance at a low and affordable price. Open to self-hosting as well.
Distributed
Citrus is being built with a distributed architecture in mind. Get the same high performance from day 1.
Coming soon

Search for anything

Use citrus to save articles, podcasts, tweets or whatever and search across them.

Retrieval-Augmented Generation (RAG) for LLMs

Provide additional context to your prompts and get even better response from GPT.

Recommendation engines

Build recommendation engines with vector embeddings and citrus.

Chatbot for your website

Index your entire website and allow users to ask questions.

Efficient vector index that fits within your stack

🍋 citrus is built to store and retrieve embeddings efficiently. It allows you to utilise your existing database so that your data always remains with you.

If you don't have one, we launch a database for you.