Gorse is a recommender system engine that combines classical machine learning and large language model rankers to generate personalized suggestions across multimodal content types.
The tool addresses the problem of building production-grade recommendation systems by providing a complete engine that handles both collaborative filtering and content-based approaches. It supports classical ranking algorithms alongside LLM-based rankers, allowing developers to choose or combine strategies based on their use case. The system processes multimodal content through embeddings, enabling recommendations across text, images, and other data types. Gorse is designed to run as a standalone service that applications can query for recommendations, abstracting away the complexity of training and serving recommendation models.
Gorse suits teams building recommendation features into applications where they want to avoid implementing recommendation logic from scratch. It works well for projects that need to handle diverse content types or want the flexibility to experiment with different ranking approaches. The tool is particularly relevant for applications requiring both traditional collaborative filtering and modern LLM-based ranking capabilities in a single system.
The project maintains active development with regular commits across its codebase. The maintainers respond to issues and pull requests, indicating ongoing engagement with the user community. Documentation is kept current alongside code changes, suggesting a commitment to keeping guidance aligned with the tool's capabilities. The project accepts external contributions and integrates them into the main codebase.
