A running catalog of apps, experiments, civic tools, learning systems, and odd little utilities I have shipped or am actively shaping. Some are serious products. Some are tests. All of them are evidence.
FeaturedProductsExperimentsCivicLearningArchive
01
experimentactive
Vellum
The visual review layer between HyperFrames and your coding agent. Pin feedback directly onto video frames so every note includes the timestamp, scene, element, coordinates, and context needed to make the edit.
Stop describing the edit. Pin it to the frame, then hand your agent the exact context.
AI VideoHyperFramesOpen Source
Adds point and region notes directly onto live HyperFrames videos.
Captures the timestamp, scene, element, coordinates, and feedback behind each edit.
Turns visual notes into Markdown and JSON instructions your coding agent can read.
Closes the loop from frame-level feedback to agent-made code changes.
AI-powered illustrated children's storybooks, on your device. Type a story idea, pick a style, and get a fully illustrated book with cover art and print-ready PDF export in minutes.
A local-first creative tool for turning story ideas into complete illustrated books.
AImacOS AppChildren's Books
Generate fully illustrated children's storybooks from a simple text prompt.
Choose from painterly, animation-style, and sketch illustration modes.
Export print-ready book formats up to 300 DPI.
Runs offline with Apple Intelligence and optional cloud upgrades.
An in-browser AI playground running 14+ models entirely on your device: speech-to-text, chat, object detection, music generation, and more, all powered by WebGPU with zero server-side processing.
A privacy-first playground for seeing what local browser AI can actually do.
AIWebGPUPrivacy-First
Runs speech, language, vision, translation, and audio models in-browser.
Keeps user data local whenever possible.
Uses WebGPU acceleration for near-native inference.
Includes object detection, segmentation, music generation, and semantic search.
An immersive AI experience that lets you have real conversations with historical figures, from Mayflower pilgrims to Jazz Age musicians to Apollo programmers.
A voice-first history experiment about presence, character, and period detail.
AIHistoryConversational
Talk with period-specific characters from different eras.
Includes Mayflower, Jazz Age, WWII, Apollo, and Wall Street scenarios.
Combines visual worldbuilding with conversational AI.
Lets visitors ask about everyday life and major historical moments.
Older creative work from broadcast, branding, websites, video, and drone projects lives here as background context, not the main act. I also keep a handful of tools and open source projects close by because the small utilities tend to explain the bigger work.
