SeaGOAT – local, “AI-based” grep for semantic code search

SeaGOAT is a code search engine for the AI age. It uses vector embeddings to enable semantic searches within your codebase. The tool requires Python 3.11 or newer and the installation of dependencies such as ripgrep and bat (optional). SeaGOAT can be used on Linux, partially on macOS, and help is needed for Windows compatibility. To use SeaGOAT, start the server with the command “seagoat-server start /path/to/your/repo” and query your repository using commands like “gt” or “seagoat”. Regular expressions can also be used in queries. SeaGOAT is licensed under an open source license and does not send data to remote servers, though future optional features might enable remote data transfer. The tool relies on vector embeddings and requires a server for optimal performance, though it can also be run locally. SeaGOAT is not a code generator, rather a code search engine, and it currently supports a variety of programming languages. The ethical implications of SeaGOAT are minimal, as it does not generate AI-derived work.

https://github.com/kantord/SeaGOAT

To top