A Model Context Protocol (MCP) server that wraps the Limadata API, enabling Claude to enrich people/company data, search for prospects, and research companies in real-time.
Perfect for: Lead scoring workflows, sales intelligence, prospect research, company deep-dives.
- enrich_person — Profile enrichment by email/LinkedIn/name+company (1-5 credits)
- enrich_company — Firmographic data by domain/LinkedIn (1 credit)
- search_people — Targeted people search by title, location, industry (2 credits)
- search_companies — Company search by criteria (2 credits)
- get_company_insights — Deep research: funding, tech stack, investors, news (5 credits)
- get_linkedin_posts — Get LinkedIn posts from profiles with pagination (1 credit per post)
- search_posts — Search LinkedIn posts by keywords and filters (2 credits)
- get_credits_balance — Check account credits (free)
- Automatic credit tracking — Response middleware captures
x-credits-*headers and appends to all outputs - Full TypeScript — Zero errors, strict mode, 100% type coverage
- Input validation — All inputs validated before API calls
- Error handling — Fast feedback on invalid input, clear API error messages
- Request history — Last 5 API calls tracked in memory
- Distributed via NPM —
npx limadata-mcpfor Claude Desktop or any MCP client
The fastest way is to use the published NPM package — no install, no clone:
Add this to your Claude Desktop config (~/Library/Application Support/Claude/claude_desktop_config.json on macOS, %APPDATA%\Claude\claude_desktop_config.json on Windows):
{
"mcpServers": {
"limadata": {
"command": "npx",
"args": ["limadata-mcp"],
"env": {
"LIMADATA_API_KEY": "your_api_key_here"
}
}
}
}Get your API key from https://app.limadata.com/settings/apikeys, then restart Claude Desktop.
git clone https://github.com/Moifek/limadata-mcp.git
cd limadata-mcp
npm install
cp .env.example .env
# Edit .env and add LIMADATA_API_KEY
npm startsrc/
├── server.ts # MCP server, tool handlers, credit formatting
├── client.ts # HTTP client, response middleware, metadata tracking
├── tools.ts # 8 tool definitions + input validators
├── types.ts # TypeScript interfaces (40+ types)
└── utils.ts # Formatting utilities
Documentation (root)
├── README.md # This file
├── QUICK_START.md # Local development setup
├── CLAUDE.md # Development guide (5-step pattern for new tools)
├── API_CONTRACT.md # Endpoint specifications
└── TESTING.md # Credit-conscious testing guide
Configuration
├── package.json
├── tsconfig.json
└── .env.example
Testing
└── test.sh # Smoke test (Credits Balance, free)
Every tool response includes a credit summary:
{...result...}
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
💳 Credits
Cost: -2
Remaining: 248
Usage: 1% of account
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
No extra configuration needed — middleware captures headers automatically.
All inputs validated before API calls:
- Prevents wasting credits on invalid requests
- Fast feedback to users
- Clear error messages
100% TypeScript, zero runtime errors:
- Full type coverage
- Strict mode enabled
- All API responses typed
- API_CONTRACT.md — Every endpoint specified
- CLAUDE.md — How to extend with new tools
npm test # Smoke test (free — Credits Balance only)
npm run build # Compile TypeScript
npm run watch # Watch mode during developmentSee TESTING.md for credit-conscious manual test snippets per tool.
The primary deployment is npx limadata-mcp via the Quick Start config above. For local development, see QUICK_START.md.
All endpoints documented in API_CONTRACT.md:
- Request schemas
- Response schemas
- Credit costs
- Rate limits
- Error codes
Want to add a new tool? See CLAUDE.md — pattern is predictable and copy-paste friendly.