Connect with MCP.

MCP is a standard way to give an AI assistant tools. Add the Mekler server in your client’s connector settings, then ask it to find a service for your task. No Mekler API key is needed for public discovery.

You are connecting for: Find recent topic news and incremental briefings with dated monitoring metadata, bounded seen-link/headline history, a ready-to-send follow-up request and source-linked headline highlights.

The connection details

Name: Mekler
Transport: Streamable HTTP

https://api.mekler.ai/mcp

This is a connection endpoint, not a web page. Your MCP client sends protocol requests to it.

Choose your client

These are client-specific configuration examples. Remote discovery is separate from the software and wallet support needed to run a service.

Claude Code

Run in your terminal, then inspect the connection with /mcp.

claude mcp add --transport http mekler https://api.mekler.ai/mcp

Claude Code connection documentation ↗

Cursor

Add this entry to your project’s .cursor/mcp.json, merging with existing servers.

{
  "mcpServers": {
    "mekler": {"url": "https://api.mekler.ai/mcp"}
  }
}

Cursor connection documentation ↗

Codex CLI

Run in the terminal where Codex is installed, then check its MCP connection.

codex mcp add mekler --url https://api.mekler.ai/mcp

Codex CLI connection documentation ↗

Python frameworks and custom agents

Use the MCP SDK in your framework’s tool adapter, or connect directly. This example uses mcp==1.30.0 and lists matching services; it does not sign or purchase anything.

python3 -m pip install "mcp==1.30.0"

import asyncio
from mcp import ClientSession
from mcp.client.streamable_http import streamable_http_client

async def main():
    async with streamable_http_client("https://api.mekler.ai/mcp") as (read, write, _):
        async with ClientSession(read, write) as session:
            await session.initialize()
            result = await session.call_tool(
                "search_services", {"query": "check a CSV file against column rules"}
            )
            print(result.model_dump_json())

asyncio.run(main())

Python MCP SDK documentation ↗

Check the connection

Ask your assistant to retrieve findTopicNews with get_service and explain its price and limits before running anything.

What this enables

Public discovery, trial instructions and referral program information are available through MCP. A discovery connection does not provide your assistant with a wallet signer or execute every paid operation. Trials need a compatible local signer; paid calls need an authorized budget and service-specific recovery.

Continue with News Search →