No trackers
The site contains no analytics scripts, ad networks, external fonts or third-party JavaScript. Everything is served from the same static directory.
Local AI Lab is a lightweight reference for people experimenting with AI on hardware they control. It focuses on deployment reality: model size, memory, quantization, workload fit and benchmark methodology.
The site contains no analytics scripts, ad networks, external fonts or third-party JavaScript. Everything is served from the same static directory.
Published benchmark scores are identified as publisher-reported. Hardware memory figures are intentionally labeled as rough estimates because runtime configuration changes the result.
There is no backend, database or build step. You can host the entire site behind a regular static web server or alongside an existing reverse-proxy setup.
The layout uses plain HTML, CSS and JavaScript. Add a page by copying an existing file, updating navigation and reusing the shared assets in assets/.
Model information on this demo site was prepared from official model announcements and documentation available in September 2026. Benchmark values can become outdated as new checkpoints, evaluation harnesses and model versions appear.