Runtime / Enforcement & Integration

GTAF SDK

Integration layer around the runtime core

GitHub repository for GTAF SDKPyPI package for GTAF SDK

GTAF SDK exists because governance artifacts and runtime contracts are not enough if every adopting team has to rebuild the same integration path from scratch.

Its role is practical: load artifacts deterministically, shape runtime context, normalize action surfaces, and wire real systems into the enforcement core without introducing semantic drift. The public implementation shows that path, but the underlying integration logic is broader than a single stack.

Public integration layer · reference implementation available

The adoption layer that helps real systems load artifacts, shape execution context, and call the runtime cleanly. The public implementation is one concrete path, but the integration model is not language-specific.

What TNT built

GTAF SDK is the adoption layer around the runtime core.

Governance models and runtime contracts are necessary, but they do not become useful at scale if every team has to rebuild loading, mapping, context shaping, and enforcement wiring on its own. The SDK exists to close exactly that gap.

Artifact loading

Deterministic loading of the runtime inputs that an enforcement decision actually depends on.

Context shaping

A practical way to normalize action surfaces and runtime context before they hit the enforcement path.

Semantic discipline

Convenience is allowed; runtime-semantic drift is not. That is the real architectural boundary of the SDK.

Why the SDK matters

Without an integration layer, governance often remains trapped between framework documentation and custom application code.

The SDK helps make runtime adoption repeatable. It standardizes the path into enforcement while making sure that the actual runtime semantics stay owned by the runtime core and do not get rewritten by helper abstractions.

Integration path

The SDK standardizes how artifacts and runtime context reach the enforcement core. It does not replace enforcement semantics; it preserves them across real integrations.

That path matters because it defines how a real product or internal system enters a governed runtime instead of bypassing it.

Current public example

The current public SDK implementation exposes the integration path concretely:

Python
from gtaf_sdk.artifacts import load_runtime_inputs
from gtaf_runtime import enforce

drc, artifacts = load_runtime_inputs(
  drc_path="path/to/drc.json",
  artifacts_dir="path/to/artifacts",
)

result = enforce(
  drc,
  context,
  artifacts,
)
A minimal integration path: artifacts are loaded first, then handed to the runtime together with context instead of being reinterpreted inside application code.

The important point is not the stack. The important point is that the integration layer remains thin enough to preserve the semantics of the runtime while still being strong enough to make adoption realistic.

How TNT can help

GTAF SDK is especially relevant for teams that already know they need runtime governance, but need a path from architecture into working systems. Typical collaboration paths include:

  • integrating governance artifacts into existing application and platform flows

  • shaping action normalization, context models, and artifact management conventions

  • defining a thin but robust contract between product code and runtime enforcement

  • avoiding the common failure mode where helper layers accidentally reinterpret policy

Where this lands first

Platform teams standardizing adoption

When multiple teams need one repeatable path into governed execution instead of rebuilding artifact handling and action mapping independently.

Products moving from prototype to control

When a promising AI system needs a cleaner operational contract between application code, governance artifacts, and the enforcement boundary beneath it.

Continue in publications

Governance & Trust Architecture Framework

GTAF Reference

A governance framework for AI systems that may act, delegate, or produce consequential effects. GTAF turns scope, authority, responsibility, and validity into structured operational artifacts.

Read the framework

Deterministic runtime enforcement core

GTAF Runtime

The enforcement core that turns evaluated governance outputs into executable allow/deny decisions. The public implementation demonstrates the contract, but the runtime model is broader than one language.

See enforcement

Infrastructure for Trusted Autonomy

ITA

A runtime architecture for systems that must act with real effect while staying governable. ITA extends GTAF into execution spaces, capability exposure, enforcement, and audit.

See the architecture

Talk to TNT when AI should do real work

From GTAF through Runtime and SDK to ITA, TNT already brings public reference work, runtime building blocks, and applied architecture into these questions. The conversation does not have to start at theory.

When these questions move from interest to implementation, TNT is a serious conversation partner.

Discuss your context