PLAYER 1 LEVEL READY
:%20window-maximize%20+%20hexagon-nodes.%20--%3e%3cpath%20fill='%23545868'%20d='M64%2064C28.7%2064%200%2092.7%200%20128L0%20384c0%2035.3%2028.7%2064%2064%2064l384%200c35.3%200%2064-28.7%2064-64l0-256c0-35.3-28.7-64-64-64L64%2064zm24%2064l336%200c13.3%200%2024%2010.7%2024%2024s-10.7%2024-24%2024L88%20176c-13.3%200-24-10.7-24-24s10.7-24%2024-24z'/%3e%3cg%20transform='translate(168.5%20200)%20scale(0.39)'%3e%3cpath%20fill='%230071ec'%20d='M248%20106.6c18.9-9%2032-28.3%2032-50.6c0-30.9-25.1-56-56-56s-56%2025.1-56%2056c0%2022.3%2013.1%2041.6%2032%2050.6l0%2098.8c-2.8%201.3-5.5%202.9-8%204.7l-80.1-45.8c1.6-20.8-8.6-41.6-27.9-52.8C57.2%2096%2023%20105.2%207.5%20132S1.2%20193%2028%20208.5c1.3%20.8%202.6%201.5%204%202.1l0%2090.8c-1.3%20.6-2.7%201.3-4%202.1C1.2%20319-8%20353.2%207.5%20380S57.2%20416%2084%20400.5c19.3-11.1%2029.4-32%2027.8-52.8l50.5-28.9c-11.5-11.2-19.9-25.6-23.8-41.7L88%20306.1c-2.6-1.8-5.2-3.3-8-4.7l0-90.8c2.8-1.3%205.5-2.9%208-4.7l80.1%2045.8c-.1%201.4-.2%202.8-.2%204.3c0%2022.3%2013.1%2041.6%2032%2050.6l0%2098.8c-18.9%209-32%2028.3-32%2050.6c0%2030.9%2025.1%2056%2056%2056s56-25.1%2056-56c0-22.3-13.1-41.6-32-50.6l0-98.8c2.8-1.3%205.5-2.9%208-4.7l80.1%2045.8c-1.6%2020.8%208.6%2041.6%2027.8%2052.8c26.8%2015.5%2061%206.3%2076.5-20.5s6.3-61-20.5-76.5c-1.3-.8-2.7-1.5-4-2.1l0-90.8c1.4-.6%202.7-1.3%204-2.1c26.8-15.5%2036-49.7%2020.5-76.5S390.8%2096%20364%20111.5c-19.3%2011.1-29.4%2032-27.8%2052.8l-50.6%2028.9c11.5%2011.2%2019.9%2025.6%2023.8%2041.7L360%20205.9c2.6%201.8%205.2%203.3%208%204.7l0%2090.8c-2.8%201.3-5.5%202.9-8%204.6l-80.1-45.8c.1-1.4%20.2-2.8%20.2-4.3c0-22.3-13.1-41.6-32-50.6l0-98.8z'/%3e%3c/g%3e%3c/svg%3e)
ai.mikedemo.dev
On-Device AI
On-Device AI runs a real language model entirely inside the browser. Inference happens on your own hardware via WebLLM and WebGPU — no server round-trips, and no data ever leaves your machine. It includes on-device inference demos, model comparisons, and diagnostics.
PROJECT DATA
Hosted on Spacefast
An AI model running inside the page — on-device inference with no server round-trips.
Built on WebLLM by MLC AI, WebGPU
OPEN ai.mikedemo.dev