Logs and explanations help after the fact, but they don’t prevent unauthorized or non-compliant actions from executing in the first place.
UAAL is a fintech-first AI governance control plane that runs inline with FastAPI services. It authorizes AI or agent actions in real time using policy, risk, and optional human approval. If a decision executes, immutable evidence is recorded; if required proof is missing, the action is blocked.
It’s packaged as a standard Python library and installs via pip:
pip install uaal-core
Happy to get feedback from folks running AI systems in regulated environments.