I'm focusing on building and shipping agentic apps and tools while experimenting with running local LLMs on macOS and Nvidia DGX Spark devices using a variety of coding harnesses like opencode, pi and Zed.
- Apple Voice Memos - Extract and process Apple Voice memo transcripts
- Text to Podcast - Generate a podcast with local AI from any text source
- Python Project Setup - Python project setup distilled into an agent skill (used for KISS Resume below)
- Agent Demo - A minimal LangChain implementation.
- Herald DNS-SD / Bonjour browser for iOS / heraldapp.app
- KISS Résumé 💋 - Generate simple, good lookling and accessible PDFs of your résumé stored as markdown.
- TabDown.app - A Safari extension to export open Safari tabs to a markdown file with Apple Intelligence AI summaries.
- vlLM & Docker on the DXG Spark - research and recipes for running the cutting edge of
vllmand open-weight models for coding on a GB10 device (DGX Spark). - llama.cpp & Docker on the DXK Spark - research and recipes for running the cutting edge of
llama.cppand open-weight models for coding on a GB10 device (DGX Spark). - Hermes-Docker - portable Hermes to run in a home lab (on a NAS).
- Chatt on Lightsail - end to end setup and maintenance steps for running chatto.run on a small, cheap light sail server.