Building the authorization layer for AI agents

Sanction Local: from runtime to install

The air-gapped runtime is real — local models, fail-closed egress denial, every denied attempt in the audit trail (see the changelog). Now the install package around it: the no-egress policy pack, and the evidence export an assessor reads without asking you to explain it. Regulated practices first.

Distribution by channel

Compatibility badges, channel-shaped policy packs, and install paths for MCP hosts, coding agents, LLM gateways, agencies, and payment-agent pilots — each previewable against your real history before you apply it.

Governed in every runtime

One authorization plane, wherever the agent runs — MCP, the SDK, and drop-in adapter paths for the frameworks agents are actually built on. Add Sanction in front of any tool server.

Sequential simulation

Today's what-if holds recorded state constant; next it replays the week in order — an early simulated denial frees budget for the request that came after, exactly as it would have lived.

Framework adapters

Drop-in hooks for LangChain, CrewAI, Vercel AI SDK, and LiteLLM — not just docs, but the adapter code that routes every tool call and model request through Sanction before it executes.

Tamper-evident audit exports

Hash-chained, exportable decision history — governance as cryptographic evidence.

Customer-managed keys + SOC 2

Bring-your-own encryption keys and the compliance attestations enterprises require.

Mandate authority (AP2 / x402)

Hold the mandate, not the rail — policy, consent, and audit in front of whichever agent-payment standard wins.