MCP Server

MCP lets you use DataStorm in plain language. After setup, you can ask for jobs directly in Claude or Cursor without writing API code.

Step 1

Get API key

You need your DataStorm API key before connecting MCP in any client.

Step 2

Add MCP server

Add the config below in Claude Desktop or Cursor MCP settings.

Step 3

Test with one prompt

Ask your assistant for jobs and verify returned results include real DataStorm data.

MCP command

npx -y @datastorm/mcp

Claude/Cursor MCP config

{
  "mcpServers": {
    "datastorm": {
      "command": "npx",
      "args": ["-y", "@datastorm/mcp"],
      "env": {
        "DATASTORM_API_KEY": "YOUR_DATASTORM_API_KEY",
        "DATASTORM_BASE_URL": "https://api.datastormai.com/api/v1"
      }
    }
  }
}

Try these prompts

"Find remote backend engineer jobs in the United States."
"Show the top companies with at least 20 open jobs."
"Get full details for this job ID: meta-123456"

Troubleshooting

Missing DATASTORM_API_KEY: Set DATASTORM_API_KEY inside your MCP config.
401 errors: Verify your API key and account access.
Timeout errors: Increase DATASTORM_TIMEOUT_MS or DATASTORM_RETRY_COUNT.

Looking for broader AI integration patterns? See AI Agents. Need advanced details? See the datastorm-mcp GitHub repository.