A fintech-first AI governance controlplane that authorizes agent actions in real

I built UAAL after repeatedly seeing AI systems in fintech make automated decisions that were impossible to defend during audits.

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.