Yebo Developer Documentation

Yebo: Global Authorization Infrastructure for the AI Economy. Every action your AI agent or application takes can now carry cryptographic proof that a real human authorized it, on their own device, at a specific time, under your exact enterprise policy.

🚀 Quick Start (Node.js / TypeScript)

To get started, install the SDK and initialize the client:

bash

npm install @yebo/sdk

javascript

import { Yebo } from "@yebo/sdk"

const yebo = new Yebo({ gateway: "https://..." })

const receipt = await yebo.authorize({
  intent: "pay Amazon $24.99",
  amount: 24.99,
  merchant: "amazon",
  asset: "USD",
  session_token: sessionToken,
})
// Returns: mandate_id,
// authorization_status, and integrity_tier

Getting Started

What is Yebo?

Core concepts - mandates, authorization flow, receipts, and the execution model

Quickstart - Authorize Your First Action

Authorize your first action in under 60 seconds

Approval Flow

How human authorization works - biometric verification, continuation tokens

Policy

How enterprise policies are evaluated against mandates

Yebo Concepts

Mandates, authorization proofs, receipts, sessions, and integrity scores

One-Liners

Copy-paste snippets for common authorization patterns

Cheatsheet

Quick reference card for endpoints, types, and status codes

Recipes

Step-by-step guides for real-world integration scenarios

Reference

SDK Reference

Every SDK method with types and examples

API Reference

Every REST endpoint, payloads, error codes

Guides

Framework Integrations

AI agents, MCP, autonomous tools, framework integrations

Execution Adapters

Stripe, Apple Pay, ACH, and custom adapter configuration

Enterprise Deployment

Policy rules, sessions, webhooks, audit export, and rollout

Security

Security model, what to keep secret, device attestation

⚙️ How It Works

The "Sentinel" system enforces a chain at the code level so that nothing can skip a step:

1
User Intent: Voice or text input.
2
Passkey Authentication: Secure biometric verification on device.
3
Mandate Sealing: Intent is formalized into an immutable mandate.
4
Verification: System invariants are checked. Any failure blocks execution.
5
Execution: The payment or action is performed.
6
Audit: A tamper-evident audit record is generated.

Nothing can skip any step. Sentinel enforces the chain at the code level. An agent cannot self-authorize. A mandate cannot be replayed. A receipt cannot be forged.

✨ Key Concepts

Feature	Description
Mandate	The atomic unit of authorization. Sealed, immutable, and contains the action, scope, and governing policy.
Proof of Authorization	A cryptographic token that proves a specific human authorized a specific action on their device.
Authorization Receipt	A portable, signed receipt returned after execution. Verifiable offline for audits or dispute resolution.
Continuation Tokens	Issued when human approval is required. Agents use these to resume execution after the user approves.
Verification	System invariant checks that run before every execution. Cannot be bypassed by configuration or code.

📊 Integrity Scores

Yebo maintains a behavioral trust score (0-1000) for every identity based on their authorization history. You can gate actions based on four tiers:

🥉Bronze
🥈Silver
🥇Gold
💎Platinum

📦 SDK Availability

TypeScript / JavaScript: Works in Node.js 18+, React, Next.js, Cloudflare Workers, Deno, and AI agent runtimes.
Swift (iOS / macOS): Available via Xcode Package Manager or Package.swift.

🔒Security Note: Yebo is designed so your application never handles cryptographic key material directly, keeping your security posture high and your integration simple.

🔑 Get Your API Key

API keys look like ent_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx. Keep them server-side.

Open Dashboard →

💬 Need Help?

Support: support@yebo.dev
Enterprise: enterprise@yebo.dev
Security: security@yebo.dev