# LLMVerify

Local guardrail server that screens model responses for prompt injections, leaked personal data, and hallucination risks before they reach users.

- MCP server · `type=mcp` · Claude Code, Codex, Cursor, GitHub Copilot, Windsurf (Devin Desktop), Cline, Zed, Gemini CLI · ★ 18
- Author: subodhkc
- License: not stated · policy=link-only · description only; the file stays with the author
- Source: https://github.com/subodhkc/llmverify-npm
- Updated: 2026-10-09
- HTML: https://affpapa.org/en/nevibecode/mcp/llmverify

## What it does

LLMVerify acts as a safety layer between your model and downstream consumers. It runs locally over `stdio` without external network calls or cloud API keys. The server exposes six tools: `verify_llm_content`, `assess_hallucination_risk`, `check_prompt_injection`, `check_pii`, `redact_pii`, and `get_llmverify_capabilities`.

Inspection relies on deterministic pattern matching and regular expressions. The engine flags standard formats of emails, phone numbers, payment cards, and known API keys. Because it relies on static rules rather than external fact-checking models, obfuscated attacks or novel injections can evade detection. Starting the MCP server requires Node.js version 20 or higher.

## Requirements
- Needs: Node.js (commands run via npx)
- Environment variables: `LLMVERIFY_MCP_MAX_INPUT_CHARS`, `LLMVERIFY_MCP_TIMEOUT_MS`, `LLMVERIFY_MCP_MAX_OUTPUT_ITEMS`, `LLMVERIFY_MCP_MAX_TEXT_FIELD_CHARS`, `LLMVERIFY_HOME`, `LLMVERIFY_AUDIT_DIR`, `LLMVERIFY_LOG_DIR`, `LLMVERIFY_BASELINE_DIR`, `LLMVERIFY_USAGE_FILE`, `LLMVERIFY_AUDIT_HASH_KEY` — keys and tokens: ask the human for the values, never commit them or put them in shared files.

## Install

### Claude Code (native platform)

**Platform CLI**
```bash
claude mcp add --env LLMVERIFY_MCP_MAX_INPUT_CHARS=<LLMVERIFY_MCP_MAX_INPUT_CHARS> --env LLMVERIFY_MCP_TIMEOUT_MS=<LLMVERIFY_MCP_TIMEOUT_MS> --env LLMVERIFY_MCP_MAX_OUTPUT_ITEMS=<LLMVERIFY_MCP_MAX_OUTPUT_ITEMS> --env LLMVERIFY_MCP_MAX_TEXT_FIELD_CHARS=<LLMVERIFY_MCP_MAX_TEXT_FIELD_CHARS> --env LLMVERIFY_HOME=<LLMVERIFY_HOME> --transport stdio llmverify -- npx -y llmverify
```
Replace <…> with your own values. 6 more environment variables are in the server README

### Codex

**Platform CLI**
```bash
codex mcp add llmverify --env LLMVERIFY_MCP_MAX_INPUT_CHARS=<LLMVERIFY_MCP_MAX_INPUT_CHARS> --env LLMVERIFY_MCP_TIMEOUT_MS=<LLMVERIFY_MCP_TIMEOUT_MS> --env LLMVERIFY_MCP_MAX_OUTPUT_ITEMS=<LLMVERIFY_MCP_MAX_OUTPUT_ITEMS> --env LLMVERIFY_MCP_MAX_TEXT_FIELD_CHARS=<LLMVERIFY_MCP_MAX_TEXT_FIELD_CHARS> --env LLMVERIFY_HOME=<LLMVERIFY_HOME> -- npx -y llmverify
```
Replace <…> with your own values. 6 more environment variables are in the server README

### Cursor

**Config file**
File `.cursor/mcp.json`:
```json
{
    "mcpServers": {
        "llmverify": {
            "command": "npx",
            "args": [
                "-y",
                "llmverify"
            ],
            "env": {
                "LLMVERIFY_MCP_MAX_INPUT_CHARS": "<LLMVERIFY_MCP_MAX_INPUT_CHARS>",
                "LLMVERIFY_MCP_TIMEOUT_MS": "<LLMVERIFY_MCP_TIMEOUT_MS>",
                "LLMVERIFY_MCP_MAX_OUTPUT_ITEMS": "<LLMVERIFY_MCP_MAX_OUTPUT_ITEMS>",
                "LLMVERIFY_MCP_MAX_TEXT_FIELD_CHARS": "<LLMVERIFY_MCP_MAX_TEXT_FIELD_CHARS>",
                "LLMVERIFY_HOME": "<LLMVERIFY_HOME>"
            }
        }
    }
}
```
Replace <…> with your own values. 6 more environment variables are in the server README

### GitHub Copilot

**Config file**
File `.vscode/mcp.json`:
```json
{
    "servers": {
        "llmverify": {
            "type": "stdio",
            "command": "npx",
            "args": [
                "-y",
                "llmverify"
            ],
            "env": {
                "LLMVERIFY_MCP_MAX_INPUT_CHARS": "<LLMVERIFY_MCP_MAX_INPUT_CHARS>",
                "LLMVERIFY_MCP_TIMEOUT_MS": "<LLMVERIFY_MCP_TIMEOUT_MS>",
                "LLMVERIFY_MCP_MAX_OUTPUT_ITEMS": "<LLMVERIFY_MCP_MAX_OUTPUT_ITEMS>",
                "LLMVERIFY_MCP_MAX_TEXT_FIELD_CHARS": "<LLMVERIFY_MCP_MAX_TEXT_FIELD_CHARS>",
                "LLMVERIFY_HOME": "<LLMVERIFY_HOME>"
            }
        }
    }
}
```
Replace <…> with your own values. 6 more environment variables are in the server README

### Gemini CLI

**Platform CLI**
```bash
gemini mcp add -e LLMVERIFY_MCP_MAX_INPUT_CHARS=<LLMVERIFY_MCP_MAX_INPUT_CHARS> -e LLMVERIFY_MCP_TIMEOUT_MS=<LLMVERIFY_MCP_TIMEOUT_MS> -e LLMVERIFY_MCP_MAX_OUTPUT_ITEMS=<LLMVERIFY_MCP_MAX_OUTPUT_ITEMS> -e LLMVERIFY_MCP_MAX_TEXT_FIELD_CHARS=<LLMVERIFY_MCP_MAX_TEXT_FIELD_CHARS> -e LLMVERIFY_HOME=<LLMVERIFY_HOME> llmverify npx -y llmverify
```
Replace <…> with your own values. 6 more environment variables are in the server README

## When to use it
- detect and redact personal data like emails or API keys in model outputs
- block prompt injections and malicious inputs before processing
- evaluate structural hallucination risks within generated text
- inspect model outputs locally without sending data over the network

## Example prompts
- Use LLMVerify to check this generated text for sensitive personal data
- Assess the prompt injection risk for the input stored in `query.txt`
- Redact all detected phone numbers and credentials from the response

## Similar
- [REA](https://affpapa.org/en/nevibecode/mcp/rea.md) — MCP server · ★ 386989
- [World Monitor](https://affpapa.org/en/nevibecode/mcp/world-monitor.md) — MCP server · ★ 89408
- [mcp](https://affpapa.org/en/nevibecode/mcp/kubeshark-mcp.md) — MCP server · ★ 12109
- [Robin: AI-Powered Dark Web OSINT](https://affpapa.org/en/nevibecode/mcp/robin-ai-powered-dark-web-osint.md) — MCP server · ★ 7658
- [jshookmcp](https://affpapa.org/en/nevibecode/mcp/jshookmcp.md) — MCP server · ★ 2061
- [OpenOSINT](https://affpapa.org/en/nevibecode/mcp/openosint.md) — MCP server · ★ 1802

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md