io.github.mkibrick/compshop

Official7 tools

by mkibrick · TypeScript

CompShop

Search 350+ compensation surveys by industry, region, or job title. Independent directory.

Search 350+ compensation surveys by industry, region, and job title.

Captured live from the server via tools/list.

search

Free-text search across the CompShop directory of compensation surveys. Returns matching vendors, reports, job families, and positions. Use for open-ended discovery questions like 'biotech surveys in Europe' or 'CEO compensation data'.

Parameters (2)

querystringrequired
Free-text query (job title, industry, vendor, geography, etc.)
limitinteger
Max results per group (default 5, max 15)

list_vendors_by_industry

List every CompShop vendor that publishes surveys for a given industry. Use when the user asks 'what survey publishers cover [industry]?'

Parameters (1)

industrystringrequired
Industry category. One of: general-industry, healthcare, life-sciences, tech, media, financial-services, insurance, energy, construction, retail, higher-ed, legal, nonprofit, executive, free.

list_vendors_by_region

List vendors with survey coverage in a given region. Use when the user asks 'what publishers cover [region]?' Region is matched against both vendor-level scope and individual report scopes.

Parameters (1)

regionstringrequired
One of: United States, Canada, United Kingdom, Europe, Asia Pacific, Latin America, Middle East & Africa, Global.

find_surveys_for_position

Find compensation surveys that benchmark a specific job title or position (e.g. 'Software Engineer', 'CPA', 'Sr Acct', 'head of finance'). Combines literal title matching with embedding-based semantic similarity, so abbreviations and paraphrases work — 'CPA' returns 'Certified Public Accountant', 'Sr Acct' returns 'Senior Accountant'. Reports per position are ranked to surface broad-coverage surveys (general-industry, US scope) before niche cuts.

Parameters (2)

positionstringrequired
Job title or position to look up (free text). Abbreviations and paraphrases are fine.
limitinteger
Max position matches to return (default 5, max 20)

get_vendor

Detailed info on a specific vendor by slug, including all of their reports. Use after `search` or `list_vendors_by_industry` returns a candidate.

Parameters (1)

slugstringrequired
Vendor slug (e.g. 'mercer-benchmark-database', 'pas', 'wtw').

get_report

Detailed info on a single survey report by slug.

Parameters (1)

slugstringrequired
Report slug (e.g. 'pas-aggregates-industry').

recommend_surveys

Recommend the best-fit compensation surveys given a hiring/benchmarking context. Use when the user asks 'what survey should I use for [situation]?' Returns ranked vendors with rationale. Required: industry. Optional: region, role focus.

Parameters (4)

industrystringrequired
Primary industry. One of: general-industry, healthcare, life-sciences, tech, media, financial-services, insurance, energy, construction, retail, higher-ed, legal, nonprofit, executive, free.
regionstring
Optional. One of: United States, Canada, United Kingdom, Europe, Asia Pacific, Latin America, Middle East & Africa, Global.
role_focusstring
Optional free-text describing the role types being benchmarked (e.g. 'software engineers', 'physicians', 'sales reps', 'CEO and C-suite').
limitinteger
Max recommendations (default 5, max 10)

README.md2 KBView on GitHub

CompShop

The independent directory of salary & compensation surveys. Search 350+ reports across 17+ vendors (Mercer, WTW, Aon, SullivanCotter, Gallagher, Pearl Meyer, Empsight, Culpepper, Milliman, MRA, LOMA, Birches Group, Western Management Group, and more) by job title, industry, geography, or publisher.

Tech stack

Framework: Next.js 14 (App Router) + TypeScript + Tailwind
Database: SQLite (better-sqlite3) with a file-backed DB bundled in the repo
Search: server-side FTS-ish query across vendors, reports, job families, positions, and participating orgs
SEO: ~3,500 statically-generated indexable pages (vendor / report / family / position) with JSON-LD, OG/Twitter tags, canonical URLs, and an auto-generated sitemap

Local dev

bash

npm install
npm run dev

Open http://localhost:3000.

Data pipelines

Loaders are in scripts/. Each writes to data/compshop.db.

bash

# Add/refresh a vendor (parent row)
npm run db:add-vendor <path-to-json>

# Add/refresh a single report (with optional XLSX position/participant lists)
npm run db:add-report <path-to-directory>

# Batch-load many reports at once
npm run db:add-report-batch "scripts/data/mercer-*"

See scripts/SCRAPING.md for the full data-loading workflow.

Deployment

Deploys to Vercel out of the box. The SQLite DB file (data/compshop.db) is committed to the repo and read at request time by the serverless routes; almost all pages are pre-rendered statically at build time via generateStaticParams + force-static.

Set NEXT_PUBLIC_SITE_URL in your Vercel project settings to your production domain (e.g., https://compshop.io).

Install

claude_desktop_config.json

{
  "mcpServers": {
    "compshop": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://www.comp-shop.com/api/mcp"
      ]
    }
  }
}

Desktop config is stdio-only; this bridges via mcp-remote. Native remote: Settings > Connectors.