Local-first policy diffing and CI to track breaking changes in legal documents
Deterministic, no-LLM approach to policy compliance
Every project in the order it hatched — newest first. 1256 and counting.
Local-first policy diffing and CI to track breaking changes in legal documents
Deterministic, no-LLM approach to policy compliance
GPU-accelerated Monte Carlo neutron transport for Apple Silicon
Massive speedups for nuclear simulations on Mac GPUs
GPU-accelerated cortical surface registration for brain imaging
Significant speedups via Triton kernels and PyTorch
LetPot Maker: studio for designing, previewing, and exporting 3D hydroponic accessories
Parametric 3D design with AI recipe selection
Reproducible benchmark harness with live leaderboard for planners
Captures every run, including timeouts and errors
Local assistant that downloads, organizes, and AI‑summarizes videos for playback and PPT.
AI‑powered video summarization and PPT generation.
MacOS app that bundles DeepSeek Harness for easy AI model use
Zero-config, plugin-friendly AI harness desktop
Overlay runtime that turns any responsive web app into a polished widget or extension.
Brings apps into overlays with isolation and lifecycle.
Local voice‑changing desktop app that turns ChatGPT’s speech into any persona in real time
Near‑real‑time local conversion, no external API needed
FirmForge: end‑to‑end MCU firmware verification pipeline for Arduino boards.
Real hardware compile, flash, serial verify in one tool.
AI-powered repo health analyzer that automates code review insights.
Multi-agent AI engine for deep repo analysis.
A high-performance macOS duplicate media finder built with Swift 6
Native Apple Silicon optimization and safe quarantine system
Privacy‑first ledger viewer for community funds, all in the browser.
No server, CSV only, instant filtering.
Real‑time Dice Chess server with bot API, rating ladder, and WebSocket play
Robust, scalable, fair dice engine with commit‑reveal
Encrypted storage router for PHP, routing uploads to Google Drive or local disk.
Secure, multi-backend file storage with key rotation.
Longwatch keeps Codex tasks running, auto‑retries and notifies you
Persistent Codex task manager with smart retries
AI agent CLI that runs, manages, and schedules LLM-powered tasks
All‑in‑one terminal for LLM agents, memory, cron, gateway
Animated heat‑pump schematic card for Home Assistant that visualises real‑time temperatures and flows
Real‑time, animated plumbing view
Agent Effects: test AI agents’ side‑effect reliability with fault injection
property‑based transaction testing for AI agents
LIWM: Auditable, provenance‑based memory for AI agents
Provenance‑gated memory for trustworthy agents
Offline guide for Burning Man: events, compass, friend map, on-device LLM
All offline, no server needed
Scope‑isolated AI retrieval library that keeps data from leaking across matters.
Provable isolation for vector‑store queries.
Transcribe audio, diarize speakers, and review via web UI
Fast, local Whisper + clustering, easy CLI
InlaySQL: a single-file, serverless Rust DB with native vector search
Embedded DB with vector queries