Every system.One research record.
Open runtimes exploring bounded, observable agent work.
01Complete archive
012026 / TypeScriptHyper RuntimeA deterministic runtime that authorizes proposed actions and records completion only after observed-state verification.022026 / TypeScriptDeterministic AI KernelFour practical agents where models propose the language and deterministic code owns policy, tools, and outcomes.032026 / PythonKitchen — Transparent Language RuntimeEvery model turn becomes a typed trail from input and context through action, verification, response, and memory.042026 / TypeScriptBuild GuardianA local-first lab that replays agent workflows under controlled perturbations and measures where their state begins to drift.052026 / TypeScriptAgentHub MarketplaceA marketplace where agents search and negotiate, while deterministic guardrails and human approval retain purchasing authority.062026 / PythonshovsOSA research runtime for agents that plan, use tools, remember, verify, and explain what happened through inspectable state.072026 / TypeScriptTelicA user-facing system that turns language into bounded, observable, verified work through connected deliberation and execution loops.