Independent intelligence.
Built for the real world.
We build multi-agent development setups, resilient edge network tools, and predictable visual prompt compilers. No corporate sales pitches here, just working software and research notes.
Systems & Code
Autonomous pair-programming environments, structured reasoning steering, and fast edge routing setups.
Agent Ladders & ML
Reinforcement learning agents on shared competitive grids, document OCR, and real-time game engines.
Sites & Staging
Oddball.net experimental staging playground and lightweight static web publishing at the edge.
Visual Studio
Reliable character consistency across photography, watercolor, and retro sci-fi using four-tier prompt compilers.
Recent Lab Dispatches
The Architecture of Provenance (v2.6): How Cryptographic Scaffolding Secures Agentic Systems
The authoritative technical reference for Moonglade Provenance 2.6: from in-memory database rows and Rust WebAssembly verification to Directory Merkle Registries and Byzantine atomic clock consensus.
Moonglade Provenance 2.5 is Open Source: Deterministic Verification for the Agentic Era
We have open-sourced the reference implementation of Moonglade Provenance. Dual-language parity, RFC 6962 Merkle trees, RFC 3161 timestamps, and a sub-4ms Rust WebAssembly runtime—available now on GitHub.
Token Minning: Why Brute-Force Context is the Wrong Bet for Agentic Systems
In 2025, the industry bet on million-token windows and prompt-driven evaluation. Here is why burning stochastic tokens on deterministic verification is an architectural dead end, and how Token Minning gates frontier LLM horsepower behind code and cryptographic proof.