Exa Websets MCP Server

A Model Context Protocol (MCP) server that integrates Exa's Websets API with Claude Desktop, Cursor, Windsurf, and other MCP-compatible clients.

What are Websets?

Websets are collections of web entities (companies, people, research papers) that can be automatically discovered, verified, and enriched with custom data. Think of them as smart, self-updating spreadsheets powered by AI web research.

Key capabilities:

• 🔍 Automated Search: Find entities matching natural language criteria

• 📊 Data Enrichment: Extract custom information using AI agents

• 🔄 Monitoring: Schedule automatic updates to keep collections fresh

• 🎯 Verification: AI validates that entities meet your criteria

• 🔗 Webhooks: Real-time notifications for collection updates

Available Tools

This MCP server provides the following tools:

Webset Management

Item Management

Search Operations

Enrichment Operations

Monitoring

Installation

Prerequisites

• Node.js v18 or higher

• Claude Desktop, Cursor, or another MCP-compatible client

• An Exa API key from exa.ai

Using Claude Code (Recommended)

The quickest way to set up Websets MCP:

Replace YOUR_API_KEY with your Exa API key.

Using NPX

Configuration

Claude Desktop Configuration

1. Enable Developer Mode

   • Open Claude Desktop

   • Click the menu → Enable Developer Mode

   • Go to Settings → Developer → Edit Config

2. Add to configuration file:

   macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

   Windows: %APPDATA%\Claude\claude_desktop_config.json

   { "mcpServers": { "websets": { "command": "npx", "args": [ "-y", "websets-mcp-server" ], "env": { "EXA_API_KEY": "your-api-key-here" } } } }

3. Restart Claude Desktop

   • Completely quit Claude Desktop

   • Start it again

   • Look for the 🔌 icon to verify connection

Cursor and Claude Code Configuration

Use the HTTP-based configuration:

Tool Schema Reference

⚠️ Important for AI Callers: See TOOL_SCHEMAS.md for exact parameter formats and examples.

Key Schema Rules:

• criteria must be an array of objects: [{description: "..."}] (NOT an array of strings)

• entity must be an object: {type: "company"} (NOT a string)

• options must be an array of objects: [{label: "..."}] (NOT an array of strings)

These formats ensure consistency across all tools and match the Websets API specification.

Usage Examples

Once configured, you can ask Claude to interact with Websets:

Creating a Webset

Listing and Viewing Websets

Managing Items

Setting Up Monitoring

Advanced Enrichments

Example Workflow

Here's a complete workflow for building a company research database:

1. Create the collection:

   Create a webset called "SaaS Companies" that searches for "B2B SaaS companies with $10M+ revenue"

2. Add enrichments:

   Add enrichments to extract: annual recurring revenue, number of customers, primary market segment, and tech stack used

3. Set up monitoring:

   Create a weekly monitor that searches for new companies and refreshes enrichment data for existing ones

4. View results:

   Show me all items with their enrichment data, sorted by revenue

Tool Details

create_webset

Creates a new webset collection with optional automatic population and enrichments.

Parameters:

• name (optional): Name for the webset

• description (optional): Description of what the webset contains

• externalId (optional): Your own identifier

• searchQuery (optional): Natural language query to find entities

• searchCount (optional): Number of entities to find (default: 10)

• searchCriteria (optional): Additional filtering criteria

• enrichments (optional): Array of enrichments to extract

Example:

create_enrichment

Adds a new data enrichment to extract custom information from each webset item.

Parameters:

• websetId: The ID of the webset

• description: Detailed description of what to extract

Example:

create_monitor

Sets up automated monitoring with a cron schedule.

Parameters:

• websetId: The ID of the webset

• schedule: Cron expression (e.g., "0 9 * * 1" for Mondays at 9am)

• behavior: Either "search" (find new items) or "refresh" (update existing)

• name (optional): Name for the monitor

• enabled (optional): Start enabled (default: true)

Common cron schedules:

• 0 9 * * 1 - Every Monday at 9am

• 0 0 * * * - Daily at midnight

• 0 */6 * * * - Every 6 hours

• 0 9 * * 1-5 - Weekdays at 9am

API Endpoints

The server connects to Exa's Websets API at https://api.exa.ai/v0/websets.

Full API documentation: docs.exa.ai/reference/websets

Advanced Configuration

Enable Specific Tools Only

To enable only certain tools, use the enabledTools config:

Debug Mode

Enable debug logging to troubleshoot issues:

Troubleshooting

Connection Issues

1. Verify your API key is valid

2. Ensure there are no spaces or quotes around the API key

3. Completely restart your MCP client (not just close the window)

4. Check the MCP logs for error messages

API Rate Limits

Websets API has the following limits:

• Check your plan limits at exa.ai/dashboard

• Use pagination for large websets

• Monitor API usage in your dashboard

Common Errors

• 401 Unauthorized: Invalid or missing API key

• 404 Not Found: Webset ID doesn't exist or was deleted

• 422 Unprocessable: Invalid query or criteria format

• 429 Rate Limited: Too many requests, wait and retry

Resources

• Exa Websets Documentation

• Exa Dashboard

• MCP Protocol Specification

• Get an Exa API Key

Development

Building from Source

Project Structure

License

MIT

Contributing

Contributions welcome! Please open an issue or PR at github.com/exa-labs/websets-mcp-server.

Support

• Documentation: docs.exa.ai

• Discord: Join the Exa community

• Email: support@exa.ai