Repolode
Submit
epsilla-cloud

epsilla-cloud/vectordb

Epsilla is an open-source C++ vector database for embedding similarity search, with metadata filtering, hybrid search, and LLM retrieval integrations.

Open-source alternative to:Pinecone

View on GitHubWebsite
Stars
874
Forks
45
Open issues
17
Last commit
18 days ago

What is it?

Epsilla is an open-source vector database focused on scalable, high-performance similarity search over embedding vectors. It treats vectors as just another field type inside familiar database, table, and field concepts, so developers can store structured records alongside their embeddings.

The core engine is written in C++ and uses parallel graph traversal for vector indexing. The project claims roughly 10x faster search than HNSW at over 99.9% precision. It supports metadata filtering, hybrid search that fuses dense and sparse vectors, and built-in embedding so text can be inserted and queried directly.

The database runs as a Docker container exposing a REST API, with Python, JavaScript, and Ruby clients. It also offers an experimental Python binding that runs without Docker, and integrates with LangChain and LlamaIndex. A managed cloud version is available.

Key features

  • 01Similarity search over embedding vectors at production scale.
  • 02Vectors stored as a field type within regular database tables.
  • 03Metadata filtering combined with vector queries.
  • 04Hybrid search fusing dense and sparse vectors.
  • 05Built-in embedding so natural language can be inserted and queried.
  • 06Python, JavaScript, Ruby clients and a REST API.

Use cases

  • Retrieval-augmented generation pipelines that need a vector store for documents.
  • Semantic search over text, images, or other embedded content.
  • Adding long-term memory retrieval to LLM applications.
  • Combining structured filters with similarity search in one database.

Topics

#vector-database#vector-search#embeddings#rag#semantic-search#c-plus-plus#llm#ai#chatgpt#data

FAQ

What is vectordb?
Epsilla is an open-source C++ vector database for embedding similarity search, with metadata filtering, hybrid search, and LLM retrieval integrations. Epsilla is an open-source vector database focused on scalable, high-performance similarity search over embedding vectors. It treats vectors as just another field type inside familiar database, table, and field concepts, so developers can store structured records alongside their embeddings.
Is vectordb free to use? What license does it have?
vectordb is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
Is vectordb still maintained?
Actively maintained. The last commit to vectordb was on Oct 4, 2026.
What can I use vectordb for?
Retrieval-augmented generation pipelines that need a vector store for documents. Semantic search over text, images, or other embedded content. Adding long-term memory retrieval to LLM applications. Combining structured filters with similarity search in one database.
Where can I get vectordb?
The source code and installation instructions are on GitHub: https://github.com/epsilla-cloud/vectordb

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects