The AI Ark Model Context Protocol (MCP) server enables AI-powered code editors like Cursor and Windsurf, plus general-purpose tools like Claude Desktop, to interact directly with your AI Ark API and documentation.
What is MCP?
Model Context Protocol (MCP) is an open standard that allows AI applications to securely access external data sources and tools. The AI Ark MCP server provides AI agents with:
- Direct API access to AI Ark functionality
- Real-time data from your AI Ark account
- Lead generation and enrichment for sales and marketing teams
AI Ark MCP Server Setup
AI Ark hosts a remote MCP server at https://api.ai-ark.com/v1/mcp?token={YOUR-API-KEY}. Configure your AI development tools to connect to this server. Replace {YOUR-API-KEY} with your API key from the AI Ark dashboard. To create your API key, visit your API Management Dashboard. API access is included on the Builder and Scale plans. Plans without API access (Free and Starter) return 403. Search requests return up to 100 records each.
Add to ~/.cursor/mcp.json:
{
"mcpServers": {
"ai-ark": {
"url": "https://api.ai-ark.com/v1/mcp?token={YOUR-API-KEY}"
}
}
}Add to ~/.codeium/windsurf/mcp_config.json:
{
"mcpServers": {
"ai-ark": {
"url": "https://api.ai-ark.com/v1/mcp?token={YOUR-API-KEY}"
}
}
}Claude Desktop connects to remote servers through Settings → Connectors → Add custom connector: enter the URL
https://api.ai-ark.com/v1/mcp?token={YOUR-API-KEY}.
If your plan has no Connectors, bridge the server with mcp-remote in claude_desktop_config.json (needs Node.js):
{
"mcpServers": {
"ai-ark": {
"command": "npx",
"args": ["-y", "mcp-remote", "https://api.ai-ark.com/v1/mcp?token={YOUR-API-KEY}"]
}
}
}💡 Learn how to use the AI Ark MCP server inside Claude: Watch Video
Testing Your MCP Setup
Once configured, you can test your MCP server connection:
- Open your AI editor (Cursor, Windsurf, etc.)
- Start a new chat with the AI assistant
- Ask about AI Ark - try questions like:
- "Search for companies in the technology industry in New York"
- "Find marketing managers at companies with 50-200 employees"
- "Find software engineers at Google along with their email addresses"
The AI now has access to your AI Ark account through the MCP server.
Tools
Your AI client discovers these tools by itself when it connects; you do not configure them.
| Tool | What it does |
|---|---|
people_search | Search people with person and company filters. |
company_search | Search companies, including lookalikes and filters by email platform and email security gateway. |
job_post_search | Search job posts to find companies that are hiring for a role or ask for a tool right now. |
email_finder, email_finder_results | Search people and find their verified emails as a job, then read the results. |
export_single | One person with a verified email, by AI Ark ID or LinkedIn URL. |
mobile_phone_finder | Mobile numbers of a person. |
reverse_people_lookup | The person behind an email address or a mobile number. |
personality_analysis | DISC and OCEAN personality profile with outreach advice, from a LinkedIn profile. |
industry_search, technology_search, location_search | Look up the exact filter values the search tools accept. Free. |
The tools cost the same credits as the API endpoints behind them.
Credits and rate limits in every answer
Every answer that comes from the AI Ark API ends with a usage object, so the assistant can tell you what a call cost and where your plan's limits stand. An answer on an Unlimited plan:
"usage": {
"creditsUsed": 0,
"limits": [
{ "window": "second", "limit": 10, "used": 1, "remaining": 9, "resetInSeconds": 1, "resetAt": "2026-11-06T10:42:12Z" },
{ "window": "4 hours", "limit": 1000, "used": 340, "remaining": 660, "resetInSeconds": 4668, "resetAt": "2026-11-06T12:00:00Z" },
{ "window": "7 days", "limit": 5000, "used": 910, "remaining": 4090, "resetInSeconds": 479868, "resetAt": "2026-11-12T00:00:00Z" }
]
}creditsUsedis what that one call was charged. It is0on Unlimited plans, for free calls, and when nothing was found.limitshas one entry per request limit of your plan: how many requests the window allows, how many are used and left, and when it starts over. Which windows you see depends on your plan; see Rate limits and quotas.
When a limit is reached, the answer says which one and how long to wait, and the assistant holds off until then:
{
"error": "429 API rate limit exceeded for this endpoint. All 1000 requests per 4 hours are used. Retry in 1 h 17 min (at 2026-11-06T12:00:00Z), not sooner.",
"retryAfterSeconds": 4668,
"retryAt": "2026-11-06T12:00:00Z",
"usage": { "creditsUsed": 0, "limits": [ { "window": "4 hours", "limit": 1000, "used": 1000, "remaining": 0, "resetInSeconds": 4668, "resetAt": "2026-11-06T12:00:00Z" } ] }
}