CueMap

CueMap functions as an open-source memory engine designed for autonomous programs, crafted in Rust. It offers: * Deterministic lexical and structural recall * Customized local MiniLM reranker (no external calls) * Millisecond-class read latencies * High recall accuracy on benchmark datasets * Inspectable traces and source-backed results This robust system delivers rapid and precise context retrieval, essential for applications requiring durable and consistent memory. It seamlessly integrates a unique local reranking mechanism, ensuring that relevant information is retrieved efficiently without dependence on external services, thus enhancing privacy and performance. CueMap is engineered for high-performance use cases, demonstrating exceptional recall rates across various memory evaluations. Whether deployed via its Python/TypeScript SDKs, Docker, CLI, or MCP, developers gain full transparency with inspectable traces and direct access to source-backed results. This makes it ideal for building reliable, long-running assistants, coding agents, and complex decision support systems. It is perfect for developers and engineering teams building sophisticated autonomous programs, requiring reliable, fast, and explainable context recall. Use cases include persistent memory for coding agents, long-running assistants, support and incident investigation, time-sensitive personalization, and comprehensive organizational memory management.