on-device-ai

Running models directly on phones, laptops, and browsers instead of shipping data to a cloud API. This area covers the tooling and tradeoffs of local inference: WebGPU and WebLLM in the browser, quantization to fit hardware limits, latency and privacy gains, and the constraints that come with running without a server. Expect practical builds, benchmarks, and honest notes on what works on real devices today and what still needs the cloud.