Connect GAIP and make one call.

Claude Desktop and claude.ai

Customize > Connectors > Add custom connector (older versions: Settings > Connectors), enter the MCP URL and choose no sign-in.

https://www.gaipagents.com/mcp

Format: client documentation.

Claude Code

claude mcp add --transport http gaip https://www.gaipagents.com/mcp

Or in the project's .mcp.json:

{
  "mcpServers": {
    "gaip": {
      "type": "http",
      "url": "https://www.gaipagents.com/mcp"
    }
  }
}

Format: client documentation.

Cursor · .cursor/mcp.json

{
  "mcpServers": {
    "gaip": {
      "url": "https://www.gaipagents.com/mcp"
    }
  }
}

Add to Cursor (opens the app; one-click install, format)

Format: client documentation.

VS Code · .vscode/mcp.json

{
  "servers": {
    "gaip": {
      "type": "http",
      "url": "https://www.gaipagents.com/mcp"
    }
  }
}

Install in VS Code (opens the app; one-click install, format)

Format: client documentation.

Gemini CLI

gemini extensions install https://github.com/tjcgraham-rgb/gaip-agents

Or in ~/.gemini/settings.json:

{
  "mcpServers": {
    "gaip": {
      "httpUrl": "https://www.gaipagents.com/mcp"
    }
  }
}

Format: client documentation.

Windsurf · ~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "gaip": {
      "serverUrl": "https://www.gaipagents.com/mcp"
    }
  }
}

Format: client documentation.

Cline · cline_mcp_settings.json

{
  "mcpServers": {
    "gaip": {
      "type": "streamableHttp",
      "url": "https://www.gaipagents.com/mcp"
    }
  }
}

Format: client documentation.

ChatGPT (developer mode connector)

Settings > Security and login > turn on Developer mode; then add a plugin with the plus button and, under Connection, enter the MCP URL (including /mcp).

https://www.gaipagents.com/mcp

Format: client documentation.

OpenAI Agents SDK (Python)

pip install openai-agents

Hosted MCP: OpenAI calls GAIP for the model (GAIP's tools are read-only):

from agents import Agent, HostedMCPTool, Runner

agent = Agent(name="Assistant", tools=[HostedMCPTool(tool_config={
    "type": "mcp", "server_label": 'gaip', "server_url": 'https://www.gaipagents.com/mcp', "require_approval": "never"})])
print(Runner.run_sync(agent, "Check https://mcp.deepwiki.com/mcp with gaip_check.").final_output)

Or run the MCP client in your own process:

import asyncio
from agents import Agent, Runner
from agents.mcp import MCPServerStreamableHttp


async def main():
    async with MCPServerStreamableHttp(name='gaip', params={"url": 'https://www.gaipagents.com/mcp'}) as server:
        agent = Agent(name="Assistant", instructions="Use the GAIP tools to check agents.",
                      mcp_servers=[server])
        result = await Runner.run(agent, "Check https://mcp.deepwiki.com/mcp with gaip_check.")
        print(result.final_output)


asyncio.run(main())

The agent's model needs your own OpenAI API key; GAIP needs none.

Format: client documentation.

Vercel AI SDK (TypeScript)

npm install @ai-sdk/mcp

import { createMCPClient } from '@ai-sdk/mcp';

const gaip = await createMCPClient({ transport: { type: 'http', url: 'https://www.gaipagents.com/mcp' } });
const tools = await gaip.tools();   // pass to generateText or streamText; then await gaip.close()

Format: client documentation.

LangChain (Python): langchain-gaip

pip install langchain-gaip

from langchain_gaip import get_tools

tools = get_tools()   # pass them to any LangChain or LangGraph agent
print(tools[0].invoke({"url": 'https://mcp.deepwiki.com/mcp'}))

Format: client documentation.

LangChain (Python): any MCP server

pip install langchain-mcp-adapters

import asyncio
from langchain_mcp_adapters.client import MultiServerMCPClient


async def main():
    client = MultiServerMCPClient({'gaip': {"transport": "http", "url": 'https://www.gaipagents.com/mcp'}})
    tools = await client.get_tools()
    print([tool.name for tool in tools])


asyncio.run(main())

Format: client documentation.

CrewAI (Python)

pip install crewai mcp

from crewai import Agent

agent = Agent(role="Integration checker", goal="Check agents and MCP servers before relying on them",
              backstory="Uses GAIP's dated record",
              mcps=['https://www.gaipagents.com/mcp'])

To give the agent one tool only, end the URL with #gaip_check.

Format: client documentation.

A2A (any A2A 0.3 or 1.0 client)

Point the client at GAIP's agent card; plain text works:

https://www.gaipagents.com/.well-known/agent-card.json

curl -s https://www.gaipagents.com/a2a/agentverse -H 'content-type: application/json' -d '{"jsonrpc":"2.0","id":1,"method":"message/send","params":{"message":{"kind":"message","role":"user","messageId":"m1","parts":[{"kind":"text","text":"check https://huggingface.co/mcp"}]}}}'

Format: client documentation.

x402

Nothing to set up. GAIP is free: it never answers with HTTP 402 and publishes no payment address, so it is not in the x402 Bazaar, a catalogue of payment-gated services. /.well-known/x402 says so for x402 clients. To see what an x402 seller charged on a date, call gaip_record_prices.

Format: client documentation.

Claude Code

claude mcp add --transport http gaip https://www.gaipagents.com/mcp

Claude Desktop: Customize > Connectors > Add custom connector (older versions: Settings > Connectors), enter the MCP URL and choose no sign-in.

Cursor · .cursor/mcp.json

{
  "mcpServers": {
    "gaip": {
      "url": "https://www.gaipagents.com/mcp"
    }
  }
}

Add to Cursor

VS Code · .vscode/mcp.json

{
  "servers": {
    "gaip": {
      "type": "http",
      "url": "https://www.gaipagents.com/mcp"
    }
  }
}

Install in VS Code