---
title: "Try · KYE Protocol™"
description: "Try KYE Protocol™ — KYE™ Agent Dev Kit™ withKye() one-line wrap, shadow mode, and the sandbox. Build + operate the rail before you adopt it."
url: https://kyeprotocol.com/journey/try/
lang: en
source: "KYE Protocol"
---

> Try KYE Protocol™ — KYE™ Agent Dev Kit™ withKye() one-line wrap, shadow mode, and the sandbox. Build + operate the rail before you adopt it.

Try

# Try the rail in shadow mode, zero lock-in.

Build against the rail and run it without risk — KYE™ Agent Dev Kit™ ships withKye() so every privileged action auto-emits the evidence family, and shadow mode observes every call before enforcing anything. Start in the sandbox.

Hand-authored hubs

## Curated entry points for this audience.

These are the hand-authored landings this audience usually starts from. The full reference index below lists every tagged page.

- [KYE™ Agent Dev Kit™](https://kyeprotocol.com/agent-dev-kit/) The opinionated SDK + CLI + templates — self-governing agents by default.
- [Builders hub](https://kyeprotocol.com/builders/) SDK overview, starter templates, sandbox.

Full reference index — every page in the try stage (20) ↓

Derived from the graph

## 20 pages in the try stage.

This page is a derived view over `internal` — a secondary projection over the canonical `kye-stage` enum (the buyer-journey spine). Adding or retagging a page rebuilds this landing on the next CI run — no hand-curated list to maintain.

Operate

## Operate — 6 pages

Run KYE Protocol™ in production — deploy, govern, observe, recover.

- [KYE™ Audit Pilot™ — 28-day buyer playbook](https://kyeprotocol.com/playbooks/audit-pilot/) A 28-day, shadow-mode-only KYE™ Audit Pilot™ from kickoff to signed pilot report. Concrete steps, named artefacts, regulatory-citation table, owner column, exit criteria.
- [KYE Protocol™ — Playbooks · buyer / vendor / regulator run-throughs](https://kyeprotocol.com/playbooks/) Three buyer-facing run-throughs for KYE Protocol™: how a buyer runs the 28-day Audit Pilot, how a vendor self-attests conformance, and how a regulator verifies a KYE™-anchored decision
- [KYE Protocol™ — Status · operational targets, SLA tiers, incident severity](https://kyeprotocol.com/status/sla/) Honest status spec, implementation, and the service that doesn t exist — Operational targets KYE Protocol™ services run under. SLA tiers (99.9% / 99.95% / 99.99%), RTO/RPO per component, incident seve
- [KYE™ Shadow Mode™ — Observe AI-agent actions, suppress side effects](https://kyeprotocol.com/shadow-mode/) A flag, not new code — KYE™ Shadow Mode™ — the runtime flag that suppresses side effects at the Commit Boundary while still running every Authority, Purpose and Decision Engine check
- [KYE™ conformance run — vendor playbook](https://kyeprotocol.com/playbooks/conformance-run/) A vendor playbook for running the 129-fixture KYE™ conformance pack against your gateway, generating a signed conformance-report.json, and posting the verification record to the public registry.
- [KYE™ regulator probe — verifier playbook](https://kyeprotocol.com/playbooks/regulator-probe/) A regulator / auditor / dispute-panel playbook for verifying a KYE™-anchored decision offline using only the publisher

Build

## Build — 14 pages

Build against the rail — SDKs, agents, conformance pack, sandbox.

- [Developer Quickstart · KYE Protocol™](https://kyeprotocol.com/docs/quickstart/) Developer Quickstart: send one authority-decision request, read the verdict and its Evidence Pack™ reference, then verify it offline against the JWKS — your first governed decision in under 10 minutes.
- [KYE™ Agent Dev Kit™ — build AI agents that prove every action they took](https://kyeprotocol.com/agent-dev-kit/) KYE™ Agent Dev Kit™ (constitution §32). Opinionated SDK + CLI + templates for building KYE™-aware AI agents. One-line \`withKye()\` wrapper. Every privileged action auto-evidenced
- [KYE™ GovernedUI™ — The human-control interface for AI-agent authority](https://kyeprotocol.com/governed-ui/) What is KYE™ GovernedUI™? The human-facing visible control surface for AI-agent authority. Approve, reject, edit, escalate, evidence, audit before an AI agent commits a consequential action.
- [KYE Protocol™ — Agent Library™ · signed reference agents for every regulated industry](https://kyeprotocol.com/agent-library/) KYE™ Agent Library™ — 23 versioned, signed reference agents for 12 categories spanning banking, payments, insurance, healthcare, pharma, logistics, energy, regtech, AI governance
- [KYE Protocol™ — Build with KYE™ | Runtime Authority API · SDKs · MCP · Connectors · KYE™ Signal Bus™](https://kyeprotocol.com/build/) Build authority-aware apps with KYE Protocol™. Runtime Authority API, three SDKs (TypeScript / Python / Go), KYE™ MCP Server, Connector Hub™, signed webhooks
- [KYE Protocol™ — Developer portal](https://kyeprotocol.com/developers/) how do you wire AI-agent authorisation, evidence, and replay into your stack with an open SDK in hours instead of weeks?
- [KYE Protocol™ — Interactive API Explorer](https://kyeprotocol.com/developers/api/) Explore the KYE Protocol™ v3 API interactively. Browse the Core entity hierarchy, State Machine, and State Library endpoints with live schema documentation.
- [KYE Protocol™ — KYE™ MCP Server | Authority checks, schemas, evidence packs for MCP agents](https://kyeprotocol.com/mcp/) The KYE™ MCP Server exposes KYE Protocol™ schemas, dictionaries, authority checks, Decision Maps™ and evidence packs to MCP-compatible agents, IDEs and developer tools — through read-only resources
- [KYE Protocol™ — Load KYE™ as an agent tool in one file](https://kyeprotocol.com/agent-native/) Agents are first-class principals: canonical identity, authority bindings, directory, metering. The KYE™ Agent Tool Pack™ ships every API as a signed.kye-tool.
- [KYE Protocol™ — Rules Gateway™ — runtime enforcement](https://kyeprotocol.com/rules-gateway/) Every delegated AI action passes through the Rules Gateway™™. The gateway loads the applicable rule packs + dictionaries + manifests for the tenant
- [KYE Protocol™ — Runtime gateway · step 7 of the KYE™ stack](https://kyeprotocol.com/gateway/) KYE™ Reference Gateway™ — the enforcement point that intercepts a candidate action, asks the policy engine for a verdict, and only lets the action commit when the signed Decision Map™ allows
- [KYE Protocol™ — Sandbox · build an authorize call in your browser](https://kyeprotocol.com/sandbox/) In-browser KYE Protocol™ sandbox. Build an /v1/runtime/authorize request, see the assembled JSON, copy the curl, see a deterministic stub Decision Map™ response. No installation, no signup.
- [KYE™ × Weights & Biases — govern agents by model provenance · KYE Protocol™](https://kyeprotocol.com/weights-and-biases/) How does KYE™ work with Weights & Biases? Connect your W&B registry; when an agent
- [KYE™ at the inference boundary (NVIDIA NIM · vLLM) — authority before the model runs · KYE Protocol™](https://kyeprotocol.com/nvidia-nim/) How does KYE™ govern AI at the inference boundary? A KYE™ PEP an OpenAI-compatible inference gateway (NVIDIA NIM, Triton, vLLM) consults before running the model — admit, block, escalate or quarantine.

## Ready to evaluate KYE Protocol™?

Start in shadow mode. We’ll deliver your first Evidence Pack™ in 4–8 weeks.

[Apply for a pilot →](https://kyeprotocol.com/pilot-apply/) [Try the sandbox →](https://kyeprotocol.com/sandbox/demos/)

Canonical KYE™ surfaces referenced on this page: [Audit Pilot](https://kyeprotocol.com/auditors/) · [Evidence Pack™](https://kyeprotocol.com/evidence-pack/) · [KYE™ Conformance Pack™](https://kyeprotocol.com/compliance/) · [KYE Protocol™](https://kyeprotocol.com/).