# multivon-mcp

Source: https://glimind.com/server/pypi%2Fmultivon-mcp
Registry id: `pypi/multivon-mcp`
Status: unknown

## Install

```json
{
  "mcpServers": {
    "multivon-mcp": {
      "command": "uvx",
      "args": [
        "multivon-mcp"
      ]
    }
  }
}
```

Claude Code:

```sh
claude mcp add multivon-mcp -- uvx multivon-mcp
```

## Tools (18)

- `eval_answer_accuracy`: QAG-graded semantic equivalence vs ground truth.
- `eval_audit_pack`: Build a procurement-ready ZIP with a SHA-256 file manifest from a pdfhell run.
- `eval_bias`: QAG-graded bias across gender, race, politics, age, socioeconomic axes.
- `eval_context_precision`: RAG retrieval quality — are the retrieved chunks on-topic?
- `eval_context_recall`: RAG retrieval completeness — does context contain enough info to answer?
- `eval_discover`: Full machine-readable capability catalog (evaluators, traps, suites, calibration data, versions). Call first.
- `eval_document_grounding`: Multi-page document-grounded faithfulness for document-AI agents.
- `eval_faithfulness`: QAG-graded faithfulness — is a RAG output grounded in the retrieved context?
- `eval_g_eval`: G-Eval flexible scoring.
- `eval_hallucination`: QAG-graded hallucination — does the output contain content NOT in context?
- `eval_pii_detection`: Local-only regex scan for PII (GDPR / CCPA / PIPEDA / HIPAA / DPDP packs).
- `eval_relevance`: QAG-graded answer-vs-question relevance.
- `eval_schema_compliance`: Validate an LLM output against a JSON Schema.
- `eval_tool_call_accuracy`: Deterministic agent tool-call correctness. No LLM.
- `eval_toxicity`: QAG-graded toxicity / harmful-content detection.
- `eval_vqa_faithfulness`: Image-grounded visual-QA faithfulness.
- `pdfhell_make`: Generate one adversarial PDF + its answer key.
- `pdfhell_run`: Run the pdfhell adversarial-PDF benchmark against a vision model. Returns pass rate, per-trap CIs, suite hash.

Measured by Glimind (https://glimind.com): MCP handshake + tool list only, no tool is ever called. CC BY 4.0.