wego

Why we included this project

Most embedding tooling assumes Python, so a Go shop that wants to train its own vectors usually has to bolt on a separate service or deal with cross-language glue. wego keeps that work in Go: it implements Word2Vec, GloVe, and LexVec from scratch, so you get an actual training pipeline rather than a wrapper around precomputed vectors, exposed through both a command-line interface and a Go SDK. The query and console commands let you inspect nearest neighbors and try simple vector arithmetic interactively, which is a quick way to sanity-check a newly trained model. Expect a focused, single-purpose library rather than an evolving platform; the project has not seen active development for a while, so weigh that before betting a long-term pipeline on it.