MCP Server
Use RequestBin directly in Claude Code, Cursor, Windsurf, and any MCP-compatible AI coding agent. Create bins, inspect captured requests, and replay them — plus create and configure Mock APIs — all without leaving your editor. Free on every plan — just create an API key and connect.
What is MCP?
The Model Context Protocol (MCP) is a standard for connecting AI coding agents to external tools. The RequestBin MCP server exposes your bins, interactions, replay, and Mock API capabilities as tools that AI agents can call directly during a conversation.
Installation
npm install -g @requestbin/mcp-server
Or run directly with npx (no install needed):
npx @requestbin/mcp-server
Get an API Key
- Go to requestbin.net/api-keys
- Create a new API key
- Copy it — you will need it for configuration below
Configure for Claude Code
Add to your project's .mcp.json or ~/.claude/settings.json:
{
"mcpServers": {
"requestbin": {
"command": "requestbin-mcp",
"env": {
"REQUESTBIN_API_KEY": "rb_your_key_here"
}
}
}
}Configure for Cursor
Add to Cursor Settings > MCP Servers:
{
"requestbin": {
"command": "requestbin-mcp",
"env": {
"REQUESTBIN_API_KEY": "rb_your_key_here"
}
}
}Configure for Windsurf
Add to ~/.windsurf/mcp_config.json:
{
"mcpServers": {
"requestbin": {
"command": "requestbin-mcp",
"env": {
"REQUESTBIN_API_KEY": "rb_your_key_here"
}
}
}
}Available Tools
| Tool | Description |
|---|---|
list_bins | List all your webhook bins with URLs and stats |
create_bin | Create a new webhook bin |
get_bin | Get bin details including recent interactions |
delete_bin | Delete a webhook bin |
list_interactions | List captured HTTP requests for a bin |
replay_request | Send an HTTP request to any URL (server-side replay) |
get_replay_status | Check the status of a replay job |
list_servers | List available servers for bin creation |
Mock API Tools
Create and configure Mock APIs — live URLs at {slug}.rbmock.dev that return responses you define — directly from your agent.
| Tool | Description |
|---|---|
list_mock_endpoints | List your Mock API endpoints, with URLs and current rule counts |
create_mock_endpoint | Create a Mock API endpoint. A starter rule (any request → 200 JSON) is added automatically so the URL returns something immediately |
add_mock_rule | Add a routing rule (method + path → status, headers, body) to a Mock API endpoint |
deploy_mock | Publish a Mock API's current rules so live requests see the latest configuration |
list_mock_captures | List recent requests captured by a Mock API endpoint, to confirm your integration is actually calling it |
Example Conversations
"Create a webhook bin for testing my Stripe integration"
The agent will call list_servers, then create_bin with name "Stripe Webhooks", and return the bin URL to configure in your Stripe Dashboard.
"Show me the last 5 requests to my Stripe webhook bin"
The agent will call list_bins to find your bins, then list_interactions with a limit of 5, and display the HTTP method, headers, and body of each request.
"Replay the last webhook event to my local server at localhost:3000/webhooks"
The agent will call list_interactions, then replay_request targeting your local URL, and confirm delivery with get_replay_status.
"Create a mock API that returns a 500 for POST /charges, then deploy it"
The agent will call create_mock_endpoint to get a live URL, then add_mock_rule to route POST /charges to a 500 response, then deploy_mock to publish it.