chromem-go

Why we included this project

Go teams that want retrieval-augmented generation, semantic search, or similar embedding-based features in their app, without standing up a separate database service, are the natural audience here. The library runs in-process the way SQLite does for relational data, and it has no third-party dependencies at all. If you already know Chroma, the API will look familiar, though it also adds more idiomatic Go methods. It supports several embedding providers, metadata filtering, and optional disk persistence, and within its scope the performance is strong: roughly 40 milliseconds to query 100,000 documents on a mid-range 2020 laptop. It is still in beta, with breaking changes possible before v1.0.0, so expect a little maintenance, but for single-process Go services that need lightweight semantic recall it avoids the operational overhead of a clustered deployment.