trustwise-mcp-server

trustwiseai/trustwise-mcp-server

3.3

If you are the rightful owner of trustwise-mcp-server and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to henry@mcphub.com.

The Trustwise MCP Server is a Model Context Protocol server designed to evaluate AI safety, alignment, and performance using advanced metrics.

The Trustwise MCP Server is a cutting-edge Model Context Protocol (MCP) server that offers a comprehensive suite of evaluation tools for assessing AI safety, alignment, and performance. It is specifically designed to help developers and AI tools programmatically evaluate the quality, safety, and cost of outputs generated by large language models (LLMs). By leveraging Trustwise's industry-leading metrics, users can ensure that AI-generated content is reliable, safe, and aligned with desired outcomes. The server is particularly useful for evaluating the safety and reliability of LLM responses, measuring the alignment, clarity, and helpfulness of AI-generated content, and estimating the carbon footprint and cost of model inference. It can be seamlessly integrated into AI pipelines, agents, or orchestration frameworks, providing robust evaluation capabilities to enhance AI development and deployment processes.

Features

  • Advanced evaluation tools for AI safety and alignment
  • Industry-leading metrics for assessing LLM outputs
  • Programmatic assessment of quality, safety, and cost
  • Integration with AI pipelines and orchestration frameworks
  • Estimation of carbon footprint and inference cost

Usages

usage with Claude Desktop

{
  "mcpServers": {
    "trustwise": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e",
        "TW_API_KEY",
        "ghcr.io/trustwiseai/trustwise-mcp-server:latest"
      ],
      "env": {
        "TW_API_KEY": "<YOUR_TRUSTWISE_API_KEY>"
      }
    }
  }
}

usage with Cursor

{
  "mcpServers": {
    "trustwise": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e",
        "TW_API_KEY",
        "-e",
        "TW_BASE_URL",
        "ghcr.io/trustwiseai/trustwise-mcp-server:latest"
      ],
      "env": {
        "TW_API_KEY": "<YOUR_TRUSTWISE_API_KEY>"
      }
    }
  }
}

Tools

  1. faithfulness_metric

    Evaluate the faithfulness of a response to its context

  2. answer_relevancy_metric

    Evaluate relevancy of a response to the query

  3. context_relevancy_metric

    Evaluate relevancy of context to the query

  4. pii_metric

    Detect PII in a response

  5. prompt_injection_metric

    Detect prompt injection risk

  6. summarization_metric

    Evaluate summarization quality

  7. clarity_metric

    Evaluate clarity of a response

  8. formality_metric

    Evaluate formality of a response

  9. helpfulness_metric

    Evaluate helpfulness of a response

  10. sensitivity_metric

    Evaluate sensitivity of a response

  11. simplicity_metric

    Evaluate simplicity of a response

  12. tone_metric

    Evaluate tone of a response

  13. toxicity_metric

    Evaluate toxicity of a response

  14. carbon_metric

    Estimate carbon footprint of a response

  15. cost_metric

    Estimate cost of a response