Developer Documentation

AgentBoost Integration Guide

Connect production skills to your desktop AI tools, headless backend services, or offline mobile devices in minutes.

01

Claude Desktop & Cursor (MCP Setup)

AgentBoost natively implements the Model Context Protocol (MCP). By pointing Claude Desktop to our production endpoint, Claude automatically mounts all 60+ deterministic skills into its context window.

claude_desktop_config.json
{
  "mcpServers": {
    "agentboost": {
      "command": "npx",
      "args": [
        "-y",
        "@modelcontextprotocol/server-fetch",
        "https://agentboost-seven.vercel.app/api/mcp?key=YOUR_API_KEY"
      ]
    }
  }
}
02

Headless Client API (Python / REST)

Automate tasks from external client machines, internal CRMs, or server workers. Submit input parameters and receive validated, deterministic returns in sub-5ms.

test_execution.py
import requests

# Direct headless invocation of any AgentBoost skill
response = requests.post(
    "https://agentboost-seven.vercel.app/api/v1/execute",
    headers={
        "Authorization": "Bearer sk_live_YOUR_KEY",
        "Content-Type": "application/json"
    },
    json={
        "skill": "crypto",
        "data": {"asset": "BTC", "timeframe": "7D"}
    }
)

data = response.json()
print("Strategy:", data['data']['strategy'])
print("Greeks:", data['data']['greeks'])
03

Google AI Edge Gallery (Mobile)

Run skills completely offline on iOS or Android using on-device models like Gemma 4:

  1. Open Google AI Edge Gallery on your mobile device.
  2. Navigate to Manage Skills → Import from URL.
  3. Paste any skill link formatted as:
    https://agentboost-seven.vercel.app/api/edge-gallery/skill-md?skillId=[skill-id]
  4. Execute tasks offline with zero data leakage.