Ahnlich

Ahnlich is an open source vector database that keeps its data in memory and adds semantic search to an application with very little setup. It ships as two services: a database that stores and ranks vectors, and an AI proxy that runs embedding models so developers can send raw text or images and let Ahnlich generate the vectors. The workflow on the site is four steps: connect a client, create a store with a chosen embedding model, insert data with metadata, and query by meaning with results ranked by similarity. It is aimed at developers who want retrieval inside their own stack without operating a cluster. The core is written in Rust with a lock free design and SIMD accelerated similarity search. The AI proxy runs ONNX embedding models, and teams can bring their own models and pick cosine, Euclidean or dot product similarity per store. Hybrid search combines vector similarity with structured metadata predicates for filtering, and snapshot persistence with deterministic hashing keeps data consistent across restarts and upgrades. Clients communicate over gRPC and Protocol Buffers, with SDKs for Rust, Python and Node.js and a Go example on the home page. Both servers are distributed as Docker images, and the site includes an audio search example that indexes a music library with CLAP embeddings and matches recorded clips. Ahnlich is free. The code is published on GitHub under the MIT license, and the project does not offer a hosted service or paid plans, so the only cost is the infrastructure it runs on.