Local-first AI workbench for llama.cpp models. Full chat UI with SSE streaming, multi-engine web search, sandboxed Python, self-aware source tools with diff previews, hot-loadable plugins, TaskMaster prompt decomposition, two-tier wiki RAG, and inline Chart.js/SVG applets. Runs entirely on your hardware.
-
Updated
Aug 29, 2026 - JavaScript