Multi LLM Cross-Check

A Model Control Protocol (MCP) server that allows cross-checking responses from multiple LLM providers simultaneously

python

GitHub Stars

python

Language

6 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "lior-ps-multi-llm-cross-check-mcp-server": {
      "command": "uv",
      "args": [
        "--directory",
        "/multi-llm-cross-check-mcp-server",
        "run",
        "main.py"
      ],
      "env": {
        "GEMINI_API_KEY": "...",
        "OPENAI_API_KEY": "sk-...",
        "ANTHROPIC_API_KEY": "sk-...",
        "PERPLEXITY_API_KEY": "..."
      }
    }
  }
}

The Multi LLM Cross-Check MCP Server lets you query multiple language model providers in parallel and receive consolidated results. This is useful when you want to compare responses side by side, verify consistency across providers, or leverage different models for diverse perspectives without sending separate requests to each API manually.

How to use

Set up the MCP server and configure your client to access it. The cross-check capability is available through a dedicated tool in conversations. When you ask to cross check with other LLMs, the server will query all configured providers in parallel and return each provider’s response, enabling quick comparisons and combined insights.

How to install

Prerequisites: you need Python 3.8 or higher and API keys for any LLM providers you plan to use. Install the uv runtime with a Python workflow.

# Prerequisites
python3 --version
pip --version

# Install the uv runtime (as shown in the setup flow)
pip install uv

Configuration and usage notes

Configure Claude Desktop (or your MCP client) to run the MCP server in stdio mode. The MCP server runs as a local process via uv and uses your API keys to reach the selected providers.

{
  "mcp_servers": [
    {
      "type": "stdio",
      "name": "cross_check",
      "command": "uv",
      "args": [
        "--directory",
        "/multi-llm-cross-check-mcp-server",
        "run",
        "main.py"
      ],
      "env": {
        "OPENAI_API_KEY": "your_openai_key",
        "ANTHROPIC_API_KEY": "your_anthropic_key",
        "PERPLEXITY_API_KEY": "your_perplexity_key",
        "GEMINI_API_KEY": "your_gemini_key"
      }
    }
  ]
}

Security and keys

Only provide API keys for the LLM providers you intend to use. If a key is not configured for a provider, that provider will be skipped and will not affect the response set.

Troubleshooting notes

If the server fails to start, verify that Python 3.8+ is installed, the uv runtime is installed, and all required API keys are provided. Ensure the directory path in the command matches your deployment location."

Examples

Example prompts you can use in a cross-check session include asking for a summary from all providers, requesting a bulleted comparison, or asking for each model’s reasoning behind a decision.

Tools and capabilities

The MCP server exposes a cross_check tool for conversations. When invoked, it queries all configured providers in parallel and returns each provider’s response for side‑by‑side comparison.

Available tools

cross_check

Cross-checks a prompt against multiple LLM providers in parallel and aggregates their responses.

Built by

VeilStrat

AI signals for GTM teams