A vendor-neutral desktop automation agent that connects any OpenWebUI-compatible LLM to OS-level controls — shell, filesystem, screenshots, accessibility-tree clicks, browser automation, and more — via a ReAct-style agentic loop.
A friendlier front-end for OpenWebUI built for higher-ed faculty who want the power of agentic AI — running commands, reading files, generating images and audio, and calling MCP servers — without having to be a developer.
Visualiztion of Bunch Clustered Software Module Dependency Graphs
A local-first multi-agent development harness that iterates a natural-language idea into a working software system through repeated agentic engineering cycles with dynamic team casting, real file actions, and automatic git commits.
A DTMF-controlled digital mode and talkgroup switcher for AllStarLink nodes running DVSwitch.
A secure and modular data collection and processing framework for heterogeneous Internet-of-Things (IoT) sensor networks.
An OpenAI-compatible HTTP proxy that aggregates multiple LLM providers behind a single endpoint, routing requests by provider prefix embedded in the model name.
A cross-platform prototype that exposes the operating system’s UI accessibility tree, textual descriptions, and ASCII spatial sketches through a browser-based dashboard and an MCP stdio server for AI agent integration.
A pi coding agent extension that registers OpenAI-compatible LLM endpoints as first-class providers inside the pi terminal harness.
A MakeCode extension for the BBC micro:bit that enables telemetry upload to a VarIOT ThingsBoard IoT gateway over WiFi using an ESP8266 module.
A Service-Based Reverse Engineering Portal
A serverless, browser-based integrated development environment for student coding practice and rapid instructor exercise development, supporting C++, Java, JavaScript, Python, R, SQL, and more without any software installation.
XML Document Visualizer