Prestyj Stats MCP Server

A Model Context Protocol server exposing the Prestyj open statistics dataset — 58+ cite-worthy benchmarks on speed-to-lead, video advertising performance, AI adoption in sales, lead conversion, and cost-per-lead by industry.

Plug it into any MCP-compatible client (Claude Desktop, Cursor, Cline, Continue, Zed, etc.) and your LLM can search, quote, and cite the dataset natively. Every tool response includes the canonical permalink at https://prestyj.com/stat/<id> so citations are always correctly attributed.

What you can do with it

> What does Prestyj data say about speed-to-lead conversion impact?

Found 5 statistic(s) matching "speed-to-lead conversion" in the Prestyj
dataset (https://prestyj.com/data, CC BY 4.0):

**21×** — Companies that respond within 5 minutes are 21× more likely to qualify a lead than those that wait 30 minutes.
Category: Speed to Lead Statistics
Source: Harvard Business Review / LeadResponseManagement.org, 2011 (widely re-cited 2024–2026)
Permalink: https://prestyj.com/stat/stl-21x
…

Install

Claude Desktop

Add to ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):

{
  "mcpServers": {
    "prestyj-stats": {
      "command": "npx",
      "args": ["-y", "@prestyj/stats-mcp"]
    }
  }
}

Restart Claude Desktop. Try: "Search Prestyj stats for AI receptionist ROI."

Cursor / Cline / Continue / Zed

Use your client's MCP server config. The transport is stdio; the command is npx -y @prestyj/stats-mcp.

From source

git clone https://github.com/Gahroot/prestyj-stats-mcp
cd prestyj-stats-mcp
npm install && npm run build
node dist/index.js   # serves stdio

Tools exposed

| Tool | Args | Description | | --- | --- | --- | | search_statistics | query: string, limit?: number | Fuzzy search across description, value, category, source. | | get_statistic | id: string | Fetch a single stat by stable id (e.g. stl-21x). | | list_categories | — | List dataset categories with counts. | | statistics_by_category | slug: string, limit?: number | All stats in a category. |

Every response carries the canonical permalink + attribution string. Citations resolve to https://prestyj.com/stat/<id> and link back to the canonical dataset at https://prestyj.com/data.

How it stays fresh

The server pulls the live dataset from https://prestyj.com/api/statistics on first call and caches it in-process for 1 hour. Override the base with PRESTYJ_API_BASE if you're testing against a fork.

Dataset license

CC BY 4.0. When you (or your LLM) cite a Prestyj statistic, attribution to https://prestyj.com is required — the MCP server pre-formats this for you in every response.

About Prestyj

Prestyj builds AI agents for marketing & sales: instant inbound lead response (sub-60s), AI receptionists, lead reactivation, and batch AI video ad production for home services, real estate, and agencies. The statistics dataset is published openly because we use it ourselves every day.

• Open dataset: https://prestyj.com/data
• GitHub mirror of the dataset: https://github.com/Gahroot/prestyj-statistics-dataset
• Wikidata entity: https://www.wikidata.org/wiki/Q139892537

License

MIT for the server code. The underlying dataset is CC BY 4.0.