Show HN: Semanticcache – A high-performance semantic caching library for Go (github.com)

Semanticcache is a new high-performance semantic caching library for Go that leverages vector embeddings to identify and retrieve semantically similar content efficiently. Designed to enhance applications involving large language models (LLMs), semantic search, and any scenario where understanding content similarity is critical, it supports multiple backend storage options including in-memory caches (LRU, LFU, FIFO) and Redis. This flexibility, combined with built-in OpenAI embedding model integration, allows developers to optimize caching strategies based on their workload and latency requirements. Technically, Semanticcache offers advanced features like customizable similarity metrics (cosine, Euclidean, dot product, and more), type-safe generics for keys and values, batch operations for efficient bulk data handling, and context-aware APIs to handle cancellations and timeouts. Its extensible architecture allows users to plug in custom embedding providers and storage backends, enabling seamless integration with existing infrastructure or bespoke systems. By bringing semantic understanding directly into the cache layer, Semanticcache empowers faster approximate nearest neighbor lookups and smarter cache hits, a significant boost for AI/ML pipelines involving LLM inference, conversational agents, and vector-based retrieval systems.