server agent in both modes; honest in-browser framing

server agent in both modes; honest in-browser framing

redesign: Highcharts + Apple-minimal SPA, data-API backend

redesign: Highcharts + Apple-minimal SPA, data-API backend

redesign: Highcharts + Apple-minimal SPA, data-API backend

streaming forecast + reliable narrative

streaming forecast + reliable narrative

fix narrative robustness + reuse distilled model

honest in-browser framing

warm models at startup

use prebuilt llama-cpp-python CPU wheel

Slipstream: gr.Server + Preact SPA, MiniCPM-1B agent + TimesFM 2.5

initial commit