Open Source
Reviewable applied-AI workflows for commerce, operations, and developer tools: facts before inference, deterministic gates, human approval, receipts, rollback, and public-safe synthetic demos.
Agent work as an operating surface, not a chat box
A dashboard concept for agent workflows. State, tools, context, approvals, and receipts are visible by default, so you can see what the agent knows, what it can touch, what it tried, and where a human has to decide.
Reviewable AI commerce changes
AI-proposed commerce changes with scoped authorization, policy gates, human approval, verification, execution receipts, and rollback paths.
Policy simulation for e-commerce
Test discount rules, shipping constraints, and promotional logic before they hit production. Think unit tests for business rules.
Personal knowledge base for AI
Crawl your documents, create embeddings, and search with semantic understanding. MCP server for Claude and other AI assistants.
Facts before inference for ops signals
Synthetic ops console that turns noisy launch and incident signals into facts, risks, proof gaps, and gated next steps.
AI-optimized log compression
Compress logs by 10-50x while preserving structure. Built for AI context windows using the Drain algorithm.
Distinct shipped projects and public-safe demos, indexed without merging their scopes.
Have questions or want to collaborate? Get in touch.