Memory you can verify — a trust layer between an AI coding agent and its own past: per-hit staleness verdicts, claim-level use attribution, measured effectiveness, episodic run-state. MCP-native, local-first, MIT, no cloud.
A deterministic, no-LLM-judge grading engine shared by model-drift and the crash-test platform
A deterministic, no-LLM-judge grading engine shared by model-drift and the crash-test platform
Topos personal AI engine (FastAPI): local data, sync, control plane WebSocket client
Remember all the tyre talk after Saturday’s Sprint? Sunday gave us some answers. Almost everyone went hard-front/medium-rear, and the leading Aprilias were still banging out serious lap times deep into the 20-lap race. Raul Fernandez produced a 1m58.703 on la…
