MCP Server

Connect AI assistants like Claude, Cursor, and others to Explorable Research using the Model Context Protocol (MCP).

What is MCP?

The Model Context Protocol (MCP) is an open standard that enables AI assistants to securely connect to external tools and data sources. Our MCP server allows AI assistants to create, manage, and iterate on your explorable research projects on your behalf.

Server Endpoint

https://mcp.explorableresearch.com/http

This is the only endpoint you need. It uses the Streamable HTTP transport (recommended by the MCP specification).

Prerequisites

API Key Required

You need an API key to authenticate with the MCP server. Create one here or see the API Keys guide.

Setup Guide

Choose your AI tool below and follow the configuration instructions.

Claude Code

Add the MCP server using the Claude Code CLI:

Add MCP server

claude mcp add \
  --header "x-api-key: YOUR_API_KEY" \
  --transport http \
  explorable-research \
  https://mcp.explorableresearch.com/http

Or add it manually to your Claude Code MCP settings file:

~/.claude/settings.json

{
  "mcpServers": {
    "explorable-research": {
      "type": "http",
      "url": "https://mcp.explorableresearch.com/http",
      "headers": {
        "x-api-key": "YOUR_API_KEY"
      }
    }
  }
}

Available Tools

Once connected, your AI assistant will have access to these tools:

create_project

Create a new explorable project from an ArXiv paper URL or PDF

continue_project

Continue an existing project with additional instructions

get_project

Get details of an existing project including preview URL

list_projects

List all your projects ordered by creation date

list_models

List available AI models for project generation

Usage Examples

Once the MCP server is configured, you can ask your AI assistant to:

Create a new project

"Create an explorable visualization from this ArXiv paper: https://arxiv.org/abs/1706.03762"

Iterate on an existing project

"Continue project abc123 and add interactive sliders to control the learning rate"

List your projects

"Show me my recent explorable research projects"

Verify Connection

To verify your MCP server is working, ask your AI assistant:

"List the available AI models from Explorable Research"

If configured correctly, you should see a list of available models.

Rate Limits

The MCP server enforces rate limits on project creation and modification:

create_project and continue_project: 10 requests per day
get_project, list_projects, list_models: No limit

Contact support if you need higher limits for your use case.

Troubleshooting

Authentication errors

Verify your API key is correct and not revoked
Ensure the header is named x-api-key (not api-key)
Check that your API key hasn't expired

Connection issues

Ensure your AI tool supports HTTP transport for MCP
Try restarting your AI tool after adding the configuration
Check that the URL is exactly https://mcp.explorableresearch.com/http

Tools not appearing

Some AI tools require a restart after configuration changes
Check your tool's logs for any MCP connection errors
Verify the JSON configuration is valid (no trailing commas, proper quotes)

Still having issues? Check the API Keys guide or contact support.