Human-Rent: Human-as-a-Service for OpenClaw

The world's first "Human-as-a-Service" platform for AI Agents

Enable AI agents to dispatch real human workers for physical world tasks that AI cannot perform.

🎯 Vision

Transform AI agents from "digital-only" to "hybrid intelligence" by giving them the ability to interact with the physical world through verified human workers.

┌─────────────────────────────────────────────────────────────┐
│  AI Agent: "I need to verify this address exists"          │
└────────────────────┬────────────────────────────────────────┘
                     │
                     ▼
┌─────────────────────────────────────────────────────────────┐
│  Human-Rent Skill: Finds nearby human worker                │
└────────────────────┬────────────────────────────────────────┘
                     │
                     ▼
┌─────────────────────────────────────────────────────────────┐
│  Human Worker: Goes to location, takes photo, verifies      │
└────────────────────┬────────────────────────────────────────┘
                     │
                     ▼
┌─────────────────────────────────────────────────────────────┐
│  AI Agent: Receives verified photo + notes, makes decision  │
└─────────────────────────────────────────────────────────────┘

✨ Key Features

• 🌍 Physical World Access - AI agents can interact with real environments
• ⚡ Async Function Calling - Non-blocking task dispatch (minutes to hours)
• 🎯 Geographic Matching - Find nearby humans automatically
• 🔐 Verification System - Cross-check, AI verification, manual review
• 💰 Pay-per-Task - Only pay when tasks complete successfully
• 🤖 MCP Protocol - Native OpenClaw integration

🚀 Quick Start

Installation

# Clone repository
git clone https://github.com/ZhenRobotics/openclaw-human-rent.git
cd openclaw-human-rent

# Install dependencies
npm install

# Test the system
./agents/human-rent-cli.sh test

Basic Usage

# Dispatch a task
./agents/human-rent-cli.sh dispatch "Take a photo of 123 Main Street"

# Check task status
./agents/human-rent-cli.sh status <task_id>

# List available humans
./agents/human-rent-cli.sh humans

# Get help
./agents/human-rent-cli.sh help

📖 Use Cases

1. Real Estate Due Diligence

human-rent dispatch "Inspect property at 456 Oak Ave. Check roof, foundation, water damage. Take 10+ photos."

Value: Investor agents can verify property condition before $1M+ decisions.

2. Vendor Verification

human-rent dispatch "Visit supplier warehouse, verify business license, inspect inventory, interview manager."

Value: Procurement agents can vet suppliers before bulk orders.

3. Address Verification

human-rent dispatch "Verify 789 Pine St exists and has 'Acme Corp' signage."

Value: Fraud detection agents can verify business legitimacy.

4. Product Quality Check

human-rent dispatch "Go to Best Buy, check if iPhone 15 Pro is in stock, note price and condition."

Value: Shopping agents can verify before purchasing.

🏗️ Architecture

Core Components

src/
├── types.ts              # TypeScript type definitions
├── task-dispatcher.ts    # Geographic matching & dispatch
├── human-pool.ts         # Human worker management
└── mcp-server.ts         # OpenClaw MCP protocol server

agents/
├── human-rent-cli.sh     # CLI interface
├── test-dispatch.ts      # Testing tool
├── dispatch-task.ts      # Task dispatcher
├── check-status.ts       # Status checker
└── list-humans.ts        # Human pool viewer

Task Types

| Type | Use Case | Latency | Cost | |------|----------|---------|------| | photo_verification | Take a photo | 5-15 min | $10-20 | | address_verification | Verify address | 10-20 min | $15-25 | | document_scan | Scan document | 10-20 min | $15-25 | | visual_inspection | Detailed inspection | 15-30 min | $20-40 | | voice_verification | Phone verification | 5-10 min | $10-20 | | purchase_verification | Check availability | 15-30 min | $20-40 |

💻 Developer Guide

📚 Code Examples

See examples/ directory for complete, runnable examples:

• basic-usage.ts - Simple photo verification
• due-diligence.ts - Investment verification workflow ($75, saves hours)
• ecommerce-qc.ts - Warehouse inspection before $50K order ($140, 357:1 ROI)

Run any example:

npx tsx examples/basic-usage.ts
npx tsx examples/due-diligence.ts
npx tsx examples/ecommerce-qc.ts

Async Task Pattern

import MCPServer from './src/mcp-server';

const server = new MCPServer();

// Dispatch task (non-blocking)
const response = await server.dispatchHuman({
  task_type: "photo_verification",
  location: "37.7749,-122.4194",
  instruction: "Take photo of building entrance",
  budget: "$15",
  timeout: "30min"
});

console.log(response.task_id); // "abc-123-def"

// Continue other work...
await doOtherStuff();

// Check status later
const status = await server.checkTaskStatus(response.task_id);
if (status.status === "completed") {
  console.log(status.result.photos); // ["https://..."]
}

Custom Task Requirements

const response = await server.dispatchHuman({
  task_type: "visual_inspection",
  instruction: "Inspect electrical panel",
  requirements: {
    minHumanRating: 4.5,
    requiredSkills: ['electrical_inspection'],
    requiredEquipment: ['multimeter'],
    certificationRequired: ['electrician_license']
  }
});

🎯 Strategic Value

Problem: AI's Physical World Blindness

Current AI agents can only work with digital information:

• ✅ Search the web
• ✅ Read documents
• ✅ Analyze data
• ❌ Verify physical reality
• ❌ Inspect real objects
• ❌ Interact with humans

Solution: Hybrid Intelligence

Human-Rent enables Human-in-the-Loop workflows:

Step 1: AI Analysis (fast, cheap, 85% confidence)
Step 2: Human Verification (slow, expensive, 95% confidence)
Step 3: AI Decision (based on verified data)

This makes AI agents trustworthy for high-stakes decisions in:

• 💰 Finance (due diligence)
• 🏥 Healthcare (physical inspection)
• ⚖️ Legal (document verification)
• 🛒 E-commerce (quality control)

Competitive Advantage

| Feature | AutoGPT | MetaGPT | OpenClaw + Human-Rent | |---------|---------|---------|---------------------------| | Web Search | ✅ | ✅ | ✅ | | Code Gen | ✅ | ✅ | ✅ | | File Ops | ✅ | ✅ | ✅ | | Physical Tasks | ❌ | ❌ | ✅ | | Human Verification | ❌ | ❌ | ✅ | | Real-world Proof | ❌ | ❌ | ✅ |

💰 Business Model

Revenue Streams

1. Per-task fees: $15-50/task
2. Platform commission: 20% of task value
3. Subscription: $99/month for unlimited tasks
4. Enterprise: Custom pricing for high-volume users

Market Size

• Target: 100K AI agents × 1 task/day × $15/task = $1.5M daily
• Platform revenue (20%): $300K daily = $109M annually

Unit Economics

| Item | Amount | |------|--------| | Human worker payment | $12 | | Platform fee (20%) | $3 | | Total task price | $15 | | Gross margin | 20% |

🧪 Current Status: MVP

✅ Implemented

• Async task dispatch system
• Geographic matching (mock data)
• 6 task types
• Task status tracking
• Simulated human completion
• MCP protocol interface
• CLI tools
• Full TypeScript types

🚧 Next Phase (3 months)

• Real human worker recruitment
• Mobile app for workers
• Stripe payment integration
• Webhook callbacks
• Cross-check verification
• Multi-city expansion

🔮 Future (6-12 months)

• Blockchain result verification
• AR glasses for workers
• Expert marketplace
• AI routing optimization
• Natural language task parsing

🛠️ Technical Stack

• Language: TypeScript
• Runtime: Node.js 18+
• Protocol: MCP (Model Context Protocol)
• Payment: Stripe (planned)
• Blockchain: Ethereum (planned)
• Mobile: React Native (planned)

📚 Documentation

• SKILL.md - OpenClaw Skill definition
• API Reference - Full API documentation (coming soon)
• Architecture - System design (coming soon)
• Human Guide - For human workers (coming soon)

🤝 Contributing

We welcome contributions! Areas where we need help:

• 🌍 Geographic expansion (non-US cities)
• 🔐 Security & privacy improvements
• 📱 Mobile app development
• 🧠 AI verification algorithms
• 📖 Documentation

⚠️ Important Notes

Legal

• Human workers are independent contractors (not employees)
• Platform provides marketplace only (no employment relationship)
• Workers assume liability for their actions
• Currently US-only (expanding after legal review)

Ethical

• No illegal tasks accepted
• No dangerous tasks without safety measures
• Privacy-first (no unnecessary PII collection)
• Fair wages (minimum $15/hour equivalent)
• Worker protection (can reject tasks)

Technical

• High latency: Minutes to hours (not milliseconds)
• Geographic dependency: Limited by human availability
• Cost: Much higher than API calls
• Reliability: Humans can cancel/fail

📄 License

MIT - See LICENSE file

📞 Contact

• GitHub: https://github.com/ZhenRobotics/openclaw-human-rent
• Issues: https://github.com/ZhenRobotics/openclaw-human-rent/issues
• ClawHub: https://clawhub.ai/ZhenStaff/human-rent
• Email: [email protected] (coming soon)

Built with ❤️ for the OpenClaw community

Make AI agents that can touch the physical world. 🌍🤖✨