Getting Started with HiveAgent MCP

Connect any MCP-compatible AI agent to 495 tools across 22 industry verticals in under 30 seconds. One endpoint, no API key required for basic access, USDC payments on Base L2 for paid tools.

MCP Endpoint

All platforms connect to the same URL: https://hiveagentiq.com/mcp — Protocol: JSON-RPC 2.0 over HTTP — Auth: None required for discovery. Tool execution fees paid in USDC on Base L2.

Three Steps to Your First Tool Call

Add the MCP server to your platform

Copy the one-line config for your platform below. Claude Desktop, Cursor, VS Code, and most MCP clients use a JSON configuration file.

Discover available tools

Call discover_tools or list_verticals to see all 495 available tools. Filter by vertical, capability, or cost.

Call any tool by name

Each tool is called by its MCP tool name with typed parameters. Paid tools automatically settle fees in USDC on Base L2 — no billing setup required if you provide a wallet address.

Platform Configuration Snippets

Pick your platform and paste the config. Every snippet connects to the same https://hiveagentiq.com/mcp endpoint.

Claude Desktop

Add to your claude_desktop_config.json file. On macOS: ~/Library/Application Support/Claude/claude_desktop_config.json. On Windows: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "hiveagent": {
      "url": "https://hiveagentiq.com/mcp"
    }
  }
}

Restart Claude Desktop after saving. HiveAgent tools will appear in Claude's tool list automatically.

Cursor

Open Cursor Settings → MCP → Add Server, or edit ~/.cursor/mcp.json directly:

{
  "mcpServers": {
    "hiveagent": {
      "url": "https://hiveagentiq.com/mcp",
      "transport": "http"
    }
  }
}

VS Code (Copilot MCP Extension)

With the GitHub Copilot Chat extension or any VS Code MCP extension, add to your settings.json:

{
  "mcp.servers": {
    "hiveagent": {
      "url": "https://hiveagentiq.com/mcp",
      "type": "http"
    }
  }
}

LangChain

Use the langchain-mcp-adapters package to connect HiveAgent as a tool provider:

# pip install langchain-mcp-adapters
from langchain_mcp_adapters.client import MCPClient
from langchain_mcp_adapters.tools import load_mcp_tools

client = MCPClient(
    server_url="https://hiveagentiq.com/mcp"
)

# Load all 495 tools as LangChain tools
tools = load_mcp_tools(client)

# Use with any LangChain agent
from langchain.agents import create_tool_calling_agent
agent = create_tool_calling_agent(llm, tools, prompt)

CrewAI

Connect HiveAgent as a CrewAI MCP server for your agent crews:

# pip install crewai crewai-tools
from crewai_tools import MCPServerTool

hiveagent_tools = MCPServerTool(
    server_url="https://hiveagentiq.com/mcp",
    name="hiveagent"
)

from crewai import Agent, Task, Crew

agent = Agent(
    role="Legal Research Specialist",
    goal="Research case law and draft legal documents",
    tools=[hiveagent_tools],
    verbose=True
)

AutoGen

Use HiveAgent with AutoGen's multi-agent framework:

# pip install pyautogen
from autogen import AssistantAgent, UserProxyAgent
from autogen.tools.mcp import MCPServerTool

mcp_config = {
    "server_url": "https://hiveagentiq.com/mcp",
    "transport": "streamable_http"
}

assistant = AssistantAgent(
    name="hiveagent_assistant",
    mcp_server=mcp_config
)

n8n

Add HiveAgent as an MCP server in n8n's AI Agent node:

# In n8n AI Agent node → Tools → Add MCP Server
Server URL: https://hiveagentiq.com/mcp
Transport: HTTP (Streamable)
Name: HiveAgent

# Or via n8n workflow JSON:
{
  "type": "@n8n/n8n-nodes-langchain.toolMcp",
  "parameters": {
    "serverUrl": "https://hiveagentiq.com/mcp"
  }
}

Zed Editor

Add to your Zed settings.json:

{
  "context_servers": {
    "hiveagent": {
      "url": "https://hiveagentiq.com/mcp"
    }
  }
}

Your First Tool Call

Once connected, test the integration by calling the ping tool. It returns server status and confirms connectivity:

// JSON-RPC 2.0 request
{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "ping",
    "arguments": {}
  }
}

// Response
{
  "result": {
    "status": "ok",
    "tools": 495,
    "verticals": 22,
    "version": "1.0"
  }
}

In Claude Desktop or Cursor, simply ask: "Ping the HiveAgent server" and the agent will call this tool automatically.

Tool Discovery Meta-Tool

HiveAgent exposes a powerful discover_tools meta-tool that lets agents programmatically list capabilities at runtime. This is especially useful for autonomous agents that need to self-configure based on the task at hand.

// List all tools in the Legal vertical
{
  "name": "discover_tools",
  "arguments": {
    "vertical": "legal",
    "include_pricing": true
  }
}

// Response excerpt
{
  "vertical": "legal",
  "tools": [
    {
      "name": "legal_case_intake",
      "description": "Structured intake for legal matters",
      "price_usdc": "0.05",
      "params": ["matter_type", "jurisdiction", "party_info"]
    },
    // ... 23 more tools
  ]
}

Discovery Parameters

Parameter	Type	Description
vertical	string	Filter by vertical slug (e.g., "legal", "healthcare", "defi")
include_pricing	boolean	Include USDC fee per tool call in the response
capability	string	Semantic search across tool descriptions (e.g., "document drafting")
free_only	boolean	Return only zero-cost tools

Workflow Tools

Beyond individual tool calls, HiveAgent provides workflow-level tools for orchestrating multi-step agent tasks:

Tool Name	Description
workflow_start	Begin a tracked multi-step workflow with optional escrow funding
workflow_step	Execute a named step within an active workflow
workflow_complete	Mark workflow done, release escrow, return results
agent_handoff	Transfer task context to another specialized agent
health_check	Check status of a running task or external integration
budget_check	Check remaining USDC budget for current session
list_verticals	Return all 22 verticals with tool counts and pricing tiers

// Example: Start a multi-step legal intake workflow
{
  "name": "workflow_start",
  "arguments": {
    "workflow_type": "legal_intake",
    "escrow_usdc": "10.00",
    "steps": ["case_intake", "jurisdiction_check", "demand_letter"]
  }
}

Next Steps

Now that you're connected, explore the rest of the documentation: