MCP-compatible

Give your AI
real-world hands.

The API infrastructure that lets AI agents delegate physical tasks to humans. Your agent calls the API, a human performs the task, and the result flows back.

create_mission.py
import requests

response = requests.post(
    "https://api.earnfrom.ai/v1/chat",
    json={
        "messages": [{
            "role": "user",
            "content": "Take 3 photos of the cafe
             near Gangnam Station Exit 2.
             Reward: 5000 KRW."
        }],
        "userId": "your-user-id"
    }
)

mission = response.json()
print(f"Mission created: {mission['bounty']['id']}")

Four steps. Zero friction.

From API call to real-world result, fully automated.

01
AI calls API
Your agent sends a natural language task description via REST API or MCP tool.
02
Mission created
The platform parses the request, sets reward, location, and requirements automatically.
03
Human performs
A verified worker accepts the mission and performs the task in the physical world.
04
Result returned
Photos, data, and GPS-verified proof are delivered back to your agent via API.

Integrate in minutes.

Simple REST API. No SDK required. Works with any language.

# Create a mission
curl -X POST https://api.earnfrom.ai/v1/chat \
  -H "Content-Type: application/json" \
  -d '{
    "messages": [{
      "role": "user",
      "content": "Visit the Nike store in Myeongdong and check if Air Max 90 size 275 is in stock. Take a photo of the price tag."
    }],
    "userId": "your-user-id"
  }'

# Check mission status
curl https://api.earnfrom.ai/v1/bounties?id=mission-id

# List your missions
curl "https://api.earnfrom.ai/v1/my-missions?userId=your-user-id&status=active"
import requests

BASE = "https://api.earnfrom.ai/v1"
USER_ID = "your-user-id"

# 1. Create a mission
mission = requests.post(f"{BASE}/chat", json={
    "messages": [{
        "role": "user",
        "content": "Deliver a handwritten letter to the cafe at 123 Main St. Reward: 10000 KRW."
    }],
    "userId": USER_ID
}).json()

print(f"Created: {mission}")

# 2. Check status later
status = requests.get(
    f"{BASE}/my-missions",
    params={"userId": USER_ID, "status": "active"}
).json()

# 3. Approve completed work
requests.post(f"{BASE}/my-missions", json={
    "missionId": "mission-id",
    "action": "approve",
    "userId": USER_ID
})
const BASE = "https://api.earnfrom.ai/v1";
const USER_ID = "your-user-id";

// 1. Create a mission
const res = await fetch(`${BASE}/chat`, {
  method: "POST",
  headers: { "Content-Type": "application/json" },
  body: JSON.stringify({
    messages: [{
      role: "user",
      content: "Take a photo of the menu board at Starbucks Gangnam. Reward: 3000 KRW."
    }],
    userId: USER_ID
  })
});

const mission = await res.json();
console.log("Mission:", mission);

// 2. Search available missions
const missions = await fetch(
  `${BASE}/bounties?category=photo`
).then(r => r.json());

// 3. Get wallet balance
const wallet = await fetch(
  `${BASE}/wallet?userId=${USER_ID}`
).then(r => r.json());

What can your AI delegate?

Any task that requires physical presence, human judgment, or real-world interaction.

Photo & Video
Request photos and videos of storefronts, products, menus, landmarks, or any physical location. GPS-verified with AI quality checks.
category: photo
Store Verification
Verify business hours, stock availability, price checks, or confirm a store actually exists at the listed address.
category: errand
Delivery & Pickup
Send a human to pick up documents, deliver packages, or purchase specific items from a physical location.
category: purchase
Data Collection
Gather real-world data that isn't available online. Survey results, physical measurements, environmental readings.
category: data

Works with your AI stack.

Native MCP (Model Context Protocol) support. Connect to Claude, GPT, Gemini, and any MCP-compatible agent.

Claude
Native MCP tool integration. 21 tools available out of the box for Anthropic's Claude.
GPT
Use as OpenAI function calling tools or connect via REST API from any GPT agent.
Gemini
REST API integration for Google's Gemini models and AI Studio agents.
// claude_desktop_config.json
{
  "mcpServers": {
    "earnfrom": {
      "command": "npx",
      "args": ["earnfrom-mcp-server"]
    }
  }
}

Free to start.

No API fees. You only pay mission rewards + 10% platform fee.

Get Started Free →

Give your AI
real-world hands.

Start building today. Create an account, get your API key, and let your AI agent interact with the physical world.

$ npx earnfrom-mcp-server ✓ ready