Now liveThe Skillselion MCP - thousands of ranked skills, loaded into your agent mid-task. No install.Get it →
Perf-Technology avatar

Perf Mcp

  • Updated February 19, 2026
  • Perf-Technology/perf-mcp

Perf MCP is a MCP server that fact-checks AI outputs with hallucination detection, schema validation, and auto-repair.

About

Perf MCP is a Model Context Protocol server that lets Claude Code, Cursor, and similar agents fact-check and remediate AI outputs instead of trusting them blindly. developers shipping features, APIs, or content with LLMs use it at review time to catch invented APIs, broken JSON, and schema drift before users see mistakes. The server connects over stdio via npm, requires a Perf API key, and exposes tooling oriented around hallucination detection, validation, and auto-repair so your coding agent can loop on fixes without you hand-auditing every response. It complements human code review rather than replacing it—ideal when generated configs, migrations, or marketing copy must match real contracts. Install when your workflow already depends on agent codegen and you need a guardrail layer on the Ship phase.

  • Hallucination detection on model outputs before they land in prod or docs
  • Schema validation against expected structures with actionable failures
  • Auto-repair flows to fix invalid AI-generated payloads in-agent
  • stdio npm MCP package (perf-mcp v0.1.1) with PERF_API_KEY from dashboard.withperf.pro

Perf Mcp by the numbers

  • Data as of Jul 7, 2026 (Skillselion catalog sync)
terminal
claude mcp add --env PERF_API_KEY=YOUR_PERF_API_KEY perf-mcp -- npx -y perf-mcp

Add your badge

Show developers this MCP server is listed on Skillselion. Paste this into your README.

Listed on Skillselion
Packageperf-mcp
TransportSTDIO
AuthRequired
Last updatedFebruary 19, 2026
RepositoryPerf-Technology/perf-mcp

What it does

Wire an agent to fact-check generated code and JSON before you merge or ship customer-facing copy.

Who is it for?

Best when you use agents for codegen, config, or structured content and want automated sanity checks before release.

Skip if: Skip if you only need static linters with no LLM in the loop, or anyone unwilling to register for a Perf API key.

What you get

Your agent can validate and repair outputs against Perf before you merge, publish, or answer customers.

  • Validated or repaired AI-generated structured output
  • Hallucination and schema failure signals the agent can act on
  • Repeatable pre-ship verification step in the agent workflow

By the numbers

  • npm package version 0.1.1
  • stdio transport
  • requires PERF_API_KEY
README.md

perf-mcp

Fact-check and fix AI outputs. Catches hallucinations, repairs broken JSON, corrects errors — before they reach users.

Works with Claude Code, Cursor, Windsurf, Cline, and any MCP-compatible client.

Quick Start

Add to your MCP client config:

{
  "mcpServers": {
    "perf": {
      "command": "npx",
      "args": ["-y", "perf-mcp"],
      "env": {
        "PERF_API_KEY": "pk_live_xxx"
      }
    }
  }
}

Get your API key at dashboard.withperf.pro — 200 free verifications, no credit card.

Tools

perf_verify

Detect and repair hallucinations in LLM-generated text. Uses multi-channel verification (web search, NLI models, cross-reference) — not just another LLM check.

perf_verify({ content: "The Eiffel Tower was built in 1887." })
→ Corrected: "built in 1887" → "inaugurated in 1889" (89% confidence)

perf_validate

Validate LLM-generated JSON against a schema and auto-repair violations. Fixes malformed enums, wrong types, missing fields, hallucinated properties.

perf_validate({
  content: '{"name": "John", "age": "twenty"}',
  target_schema: { type: "object", properties: { name: { type: "string" }, age: { type: "number" } } }
})
→ Rejected: /age must be number

perf_correct

General-purpose output correction. Classifies the error type and applies the right fix — hallucination, schema violation, semantic inconsistency, or instruction drift.

perf_correct({ content: "The Great Wall was built in 1950.", correction_budget: "fast" })
→ Corrected: temporal_error + factual_error detected (87% confidence)

perf_chat

Route LLM requests to the optimal model automatically. Selects between GPT-4o, Claude, Gemini, and 20+ models based on task complexity. OpenAI-compatible format.

Setup by Client

Claude Code

Add to your project's .mcp.json:

{
  "mcpServers": {
    "perf": {
      "command": "npx",
      "args": ["-y", "perf-mcp"],
      "env": {
        "PERF_API_KEY": "pk_live_xxx"
      }
    }
  }
}

Cursor

Settings → MCP → Add Server:

{
  "mcpServers": {
    "perf": {
      "command": "npx",
      "args": ["-y", "perf-mcp"],
      "env": {
        "PERF_API_KEY": "pk_live_xxx"
      }
    }
  }
}

Windsurf

Add to ~/.codeium/windsurf/mcp_config.json:

{
  "mcpServers": {
    "perf": {
      "command": "npx",
      "args": ["-y", "perf-mcp"],
      "env": {
        "PERF_API_KEY": "pk_live_xxx"
      }
    }
  }
}

Environment Variables

Variable Required Description
PERF_API_KEY Yes Your Perf API key (pk_live_xxx)
PERF_BASE_URL No Override API URL (for testing)

Pricing

Plan Credits Price
Free 200 verifications $0 (never expires)
Pro 1,000/mo $19/mo
Pay-as-you-go Unlimited $0.02/verification

1 tool call = 1 credit. Get started at dashboard.withperf.pro.

License

MIT

Recommended MCP Servers

How it compares

MCP-backed output verification service, not a local ESLint skill or generic chat prompt.

FAQ

Who is Perf MCP for?

Developers who ship with AI coding agents and need a second layer that catches hallucinations and schema errors before release.

When should I use Perf MCP?

Use it during Ship review whenever an agent drafts JSON, API shapes, migrations, or user-facing claims you cannot afford to get wrong.

How do I add Perf MCP to my agent?

Install the npm package perf-mcp, set PERF_API_KEY from dashboard.withperf.pro, and register the stdio server in your agent’s MCP config.

AI & LLM Toolsllmautomation

This week in AI coding

Five minutes, every Monday - the tools, releases and tactics for developers.

unsubscribe anytime.