Policy checks, evidence bundles, immutable audit trails, memory-backed orchestration, and quorum-style review patterns inside execution.
pip install ioa-coreTask, policy context, and model/provider choices arrive together.
Checks run before the action proceeds, not only after output exists.
Provider-neutral and quorum-style review patterns can participate in the decision.
The action leaves behind inspectable records instead of forcing reconstruction later.
The page needs to answer one practical question fast: what do I actually get if I wire this in? The answer is a governed execution path with concrete artifacts you can inspect, persist, and route upstream.
from ioa_core.governance import PolicyEngine
engine = PolicyEngine()
result = engine.validate_action({
"action_type": "clinical_summary",
"jurisdiction": "EU",
"contains_sensitive_data": True,
})
print(result.decision)
print(result.audit_id){
"policy": "demo-governed",
"decision": "allow_with_mitigation",
"audit_chain_verified": true,
"system_laws_applied": ["Law 1", "Law 5", "Law 7"],
"evidence_id": "ev-0001"
}Yes. IOA Core still carries the seven-law governance framing in code, docs, and evidence output. They should support the runtime narrative, not replace it.
They make the governance layer legible and give teams a stable vocabulary for policy, fairness, oversight, and sustainability checks.
Developers adopt on concrete behavior. The laws are more persuasive once the page shows what enforcement and evidence actually look like.
Read the System Laws page if you want the full framing behind the runtime controls.
Most AI tooling focuses on prompts, outputs, tracing, or post-hoc monitoring. IOA Core focuses on the runtime path itself: what policy was applied, what evidence was produced, and how a decision can be inspected later.
Put governance checks in the execution path instead of relying only on dashboards and after-the-fact review.
Generate evidence artifacts and audit records as part of governed actions, not as reconstructed summaries later.
Use provider-neutral and quorum-oriented patterns instead of locking governance to one model vendor.
IOA Core is most useful when AI workflows affect compliance, safety, privacy, or material decisions.
The current release line is live, installable, and aligned across package, docs, and website surfaces.
`ioa-core` `v2.6.1` is published on PyPI and verified from a fresh install path.
Core GitHub Actions are green and the public docs were updated to match the shipped release.
Install it, run the examples, inspect the audit and evidence outputs, and give technical feedback.
Use the package, inspect the repo, and read the docs. That is a stronger developer experience than a pure marketing pitch.