P12 dSpark runtime proof (SGLang + LFM2.5)
Serve an open model for the nexus-cortex harness
Run an in‑browser LLM test with WebGPU and see performance metrics
Generate chat responses with a customizable language model