local AI inference and runtimes
Recent window: last 5.3 months (2026-04-28 → 2026-10-05), compared with the prior 5.3 months.
These products provide lightweight inference engines, specialized runtimes, and optimization tools designed to execute machine learning models locally or on constrained hardware. They are used by machine learning engineers and systems developers seeking low-latency execution, memory efficiency, and hardware acceleration on devices like Apple Silicon or edge NPUs. This cluster is distinguished by its focus on low-level compute runtimes and kernel optimizations rather than cloud-hosted APIs or application-layer wrappers.