LeadSpark MCP
# LeadSpark MCP
Research any company or person from your AI IDE. Get company info, tech stack, contacts, and emails - all without leaving Cursor, Claude Code, or Windsurf.
## What it does
LeadSpark is an MCP server that gives your AI agent the ability to research companies and find contacts. Type "research stripe.com" and get back structured company data from 6+ free data sources.
### Tools
| Tool | Description | Data returned |
|------|------------|---------------|
| `research_company` | Full company enrichment | Name, description, industry, HQ, founded year, employees, tech stack, social links, email provider, hosting |
| `find_contacts` | Find people at a company | Names, titles, LinkedIn URLs, probable email addresses |
| `lookup_tech_stack` | Detect website technologies | CMS, frameworks, analytics, CDN, hosting, chat tools, payments (60+ fingerprints) |
### Data sources (all free)
- Website scraping (meta tags, Schema.org JSON-LD, about pages)
- DNS/MX records (email provider, hosting detection)
- SSL certificates (organization name, location)
- Built-in tech fingerprints (60+ technologies)
- GitHub API (org data, team members, languages)
- DuckDuckGo Instant Answers (Wikipedia summaries, company facts)
- Email pattern detection (first.last@, flast@, etc.)
## Setup
### Claude Code
Add to your MCP config (`~/.claude.json` or project settings):
```json
{
"mcpServers": {
"leadspark": {
"command": "npx",
"args": ["-y", "leadspark-mcp"]
}
}
}
```
### Cursor
Add to `.cursor/mcp.json`:
```json
{
"mcpServers": {
"leadspark": {
"command": "npx",
"args": ["-y", "leadspark-mcp"]
}
}
}
```
### Windsurf
Add to MCP settings:
```json
{
"mcpServers": {
"leadspark": {
"command": "npx",
"args": ["-y", "leadspark-mcp"]
}
}
}
```
### Optional: Set environment variables for better results
```bash
# Higher GitHub API rate limits (5K -> 15K requests/hour)
export GITHUB_TOKEN=your_github_token
# Hunter.io for email verification (50 free lookups/month)
export HUNTER_API_KEY=your_hunter_key
```
## Example output
```
> research stripe.com
# Stripe, Inc
**Description:** Stripe is a financial services platform...
## Company Info
- Industry: Financial Services
- Headquarters: South San Francisco, California, US
- Founded: 2011
- Employees: 3029+ GitHub followers
- Email Provider: Google Workspace
- Hosting: AWS
## Social Links
- GitHub: https://github.com/stripe
## Tech Stack
- Framework: React, Next.js
- Analytics: Google Analytics
- Payments: Stripe
```
## Development
```bash
git clone https://github.com/yourusername/leadspark-mcp
cd leadspark-mcp
npm install
npm run dev
```
## License
MIT
TDQS
Scored across 3 tools
research_company and lookup_tech_stack overlap significantly because research_company already returns tech stack, hosting, and email provider information. find_contacts is clearly distinct, but an agent could easily hesitate between the two company-research tools.
All tool names follow a consistent verb_noun pattern using lowercase snake_case: research_company, find_contacts, lookup_tech_stack. There is no mixing of conventions or vague verbs.
Three tools is a reasonable size for a focused lead-generation server. However, lookup_tech_stack feels somewhat redundant with research_company, so the count is appropriate but not perfectly scoped.
The core lead-generation workflow is covered: research a company, find contacts, and identify technologies. Obvious gaps remain, such as email verification, lead list management, or exporting results, and the tech-stack lookup is already partially included in research_company.