Fiddler MCP Server

The Fiddler Everywhere MCP (Model Context Protocol) server lets you connect coding assistants and agent-powered tools, reaching new levels of developer productivity. The MCP server provides proprietary context to agent-powered IDEs, apps, and tools. You can use the Fiddler Everywhere MCP server for comprehensive analysis of captured HTTPS traffic, including web debugging, security highlights, performance reports, data extraction, and general usage. The Fiddler MCP enables you to successfully prompt more complex questions and tasks, and generate tailored code based on the information received from the captured HTTPS traffic.

Prerequisites

To use the Fiddler Everywhere MCP server, you need:

• The latest version of Fiddler Everywhere.
• A Fiddler Everywhere Pro or higher subscription tier.
• An active internet connection.
• A unique API Key for the Fiddler Everywhere MCP server (generated within the application).
• A compatible MCP client (IDE, code editor, or app) that supports MCP tools.

Installation

The fastest way to configure the Fiddler MCP server in any supported coding tool is to use the fiddler-mcp-setup agent skill. It auto-discovers the port, retrieves your API key, writes the correct config file for your tool, and gitignores it — all in one step. If Fiddler Everywhere is not yet installed, start with the fiddler-download-setup skill instead.

To configure the MCP server manually, use the generic settings below or follow the per-tool instructions:

The generic settings for the Fiddler Everywhere MCP server are:

• Server name handle: #fiddler (default value - can be customized through Settings > MCP Server)
• Server type: http
• Server URL: http://localhost:8868/mcp (default value - the port can be customized through Settings > MCP Server)
• Authorization: A generated unique API key to be placed as an Authorization header.

GitHub Copilot in Visual Studio Code

To enable the Fiddler Everywhere MCP server for GitHub Copilot in Visual Studio Code, follow these steps:

1. Start the Fiddler Everywhere application.

2. Go to Settings > MCP Server and complete the following:

   • Set the MCP server port (default value is 8868).
   • Generate a unique API Key.
   • Copy the generated JSON configuration from the VSCode tab.

3. Open your Visual Studio Code workspace and complete the following:

   • Create a .vscode/mcp.json file in your workspace.
   • Paste the copied Fiddler Everywhere MCP configuration. Ensure that the JSON is properly formatted and that the Fiddler MCP server is within the servers property.

   example mcp.json file in VSCode with Fiddler Everywhere MCP server

   JSON

   {
       "servers": {
           "fiddler": {
               "type": "http",
               "url": "http://localhost:8868/mcp",
               "headers": {
                   "Authorization": "ApiKey FIDDLER_API_KEY_HERE"
               }
           }
       }
   }

   Replace FIDDLER_API_KEY_HERE with the API key generated in step 2.

4. Add .vscode/mcp.json to your .gitignore to avoid committing the API key to source control.

5. Start the Fiddler Everywhere MCP server in Visual Studio Code.

Cursor

To enable the Fiddler Everywhere MCP server in Cursor, follow these steps:

1. Start the Fiddler Everywhere application.

2. Go to Settings > MCP Server and complete the following:

   • Set the MCP server port (default value is 8868).
   • Generate a unique API Key.
   • Copy the generated JSON configuration from the Cursor tab.

3. Open your Cursor workspace and complete the following:

   • Create a .cursor/mcp.json file in your workspace.
   • Paste the copied Fiddler Everywhere MCP configuration. Ensure that the JSON is properly formatted and that the Fiddler MCP server is within the mcpServers property.

   example mcp.json file in Cursor with Fiddler Everywhere MCP server

   JSON

   {
       "mcpServers": {
           "fiddler": {
           "url": "http://localhost:8868/mcp",
               "headers": {
                   "Authorization": "ApiKey FIDDLER_API_KEY_HERE"
               }
           }
       }
   }

4. Start the Fiddler Everywhere MCP server in Cursor.

Claude Code

To enable the Fiddler Everywhere MCP server in Claude Code, follow these steps:

1. Start the Fiddler Everywhere application.

2. Go to Settings > MCP Server and complete the following:

   • Set the MCP server port (default value is 8868).
   • Generate a unique API Key.

3. In your project directory, create an .mcp.json file with the following configuration:

   example .mcp.json file for Claude Code with Fiddler Everywhere MCP server

   JSON

   {
       "mcpServers": {
           "fiddler": {
               "type": "http",
               "url": "http://localhost:8868/mcp",
               "headers": {
                   "Authorization": "ApiKey FIDDLER_API_KEY_HERE"
               }
           }
       }
   }

   Replace FIDDLER_API_KEY_HERE with the API key generated in step 2.

4. Add .mcp.json to your .gitignore to avoid committing the API key to source control.

5. Start Claude Code from your project directory. The Fiddler MCP server will be available automatically.

Alternatively, you can register the server via the CLI: claude mcp add --transport http fiddler http://localhost:8868/mcp -H "Authorization: ApiKey FIDDLER_API_KEY_HERE"

Claude Desktop

To enable the Fiddler Everywhere MCP server in Claude Desktop, follow these steps:

1. Start the Fiddler Everywhere application.

2. Go to Settings > MCP Server and complete the following:

   • Set the MCP server port (default value is 8868).
   • Generate a unique API Key.

3. Open the Claude Desktop configuration file for your operating system:

   • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
   • Windows: %APPDATA%\Claude\claude_desktop_config.json

4. Add the Fiddler Everywhere MCP server entry to the mcpServers object. Create the file if it does not exist.

   example claude_desktop_config.json file with Fiddler Everywhere MCP server

   JSON

   {
       "mcpServers": {
           "fiddler": {
               "type": "http",
               "url": "http://localhost:8868/mcp",
               "headers": {
                   "Authorization": "ApiKey FIDDLER_API_KEY_HERE"
               }
           }
       }
   }

   Replace FIDDLER_API_KEY_HERE with the API key generated in step 2.

5. Restart Claude Desktop to apply the configuration.

Windsurf

To enable the Fiddler Everywhere MCP server in Windsurf, follow these steps:

1. Start the Fiddler Everywhere application.

2. Go to Settings > MCP Server and complete the following:

   • Set the MCP server port (default value is 8868).
   • Generate a unique API Key.

3. Open the Windsurf MCP configuration file located at:

   • macOS / Linux: ~/.codeium/windsurf/mcp_config.json
   • Windows: %USERPROFILE%\.codeium\windsurf\mcp_config.json

4. Add the Fiddler Everywhere MCP server entry to the mcpServers object. Create the file if it does not exist.

   example mcp_config.json file in Windsurf with Fiddler Everywhere MCP server

   JSON

   {
       "mcpServers": {
           "fiddler": {
               "serverType": "http",
               "url": "http://localhost:8868/mcp",
               "headers": {
                   "Authorization": "ApiKey FIDDLER_API_KEY_HERE"
               }
           }
       }
   }

   Replace FIDDLER_API_KEY_HERE with the API key generated in step 2.

5. Restart Windsurf or reload the MCP configuration to apply the changes.

GitHub Copilot CLI

To enable the Fiddler Everywhere MCP server in GitHub Copilot CLI, follow these steps:

1. Start the Fiddler Everywhere application.

2. Go to Settings > MCP Server and complete the following:

   • Set the MCP server port (default value is 8868).
   • Generate a unique API Key.

3. Open or create the GitHub Copilot CLI MCP configuration file at ~/.copilot/mcp-config.json.

4. Add the Fiddler Everywhere MCP server entry to the mcpServers object.

   example mcp-config.json for GitHub Copilot CLI with Fiddler Everywhere MCP server

   JSON

   {
       "mcpServers": {
           "fiddler": {
               "type": "http",
               "url": "http://localhost:8868/mcp",
               "headers": {
                   "Authorization": "ApiKey FIDDLER_API_KEY_HERE"
               },
               "tools": ["*"]
           }
       }
   }

   Replace FIDDLER_API_KEY_HERE with the API key generated in step 2.

The "tools": ["*"] property is required for GitHub Copilot CLI — omitting it disables all MCP tools.

The configuration file ~/.copilot/mcp-config.json is a global user-level file stored outside any repository. Keep it private and do not share or commit it, as it contains your API key in plaintext.

Usage

To use the Fiddler Everywhere MCP server:

1. Start the Fiddler Everywhere application.
2. Configure the MCP server in your IDE:
   • GitHub Copilot in Visual Studio Code
   • GitHub Copilot CLI
   • Cursor
   • Claude Code
   • Claude Desktop
   • Windsurf
   • Or any other compatible tool that supports MCP servers.
3. Enable the agent mode in your IDE.
4. Start your prompt with #fiddler (or with # followed by your custom server name, if configured).

The MCP server will provide context from your captured HTTPS traffic to enhance your MCP-assisted development workflow.

The results and quality of outputs from the Fiddler Everywhere MCP server may vary depending on the prompt formulation and the specific language model being used. Different models have varying capabilities, and prompt quality significantly impacts the accuracy and relevance of generated responses.

You can jump start your MCP journey by exploring our prompt library or check the agent skills that automate common setup and analysis tasks.

Supported Capturing Modes

The Fiddler Everywhere MCP server supports the following capturing modes:

• Browser capturing mode
• Terminal capturing mode
• Reverse proxy

Available MCP Tools

The Fiddler Everywhere MCP server exposes the following tools. These tools can be invoked by your coding assistant (using #fiddler or your configured server name) to interact programmatically with the Fiddler Everywhere application.

Authentication and Status

Traffic Capture

Session Management

The session management tools support both the Live Traffic and Agent Calls tabs through a required sessionsSource parameter. The parameter accepts two values: LiveTraffic and AgentCalls. Use LiveTraffic for real-time captured HTTP/HTTPS traffic. Use AgentCalls for LLM/AI agent API calls. When prompting your coding assistant, explicitly specify which sessions source to target so that the sessionsSource parameter is set correctly.

Agent Cache

The Agent Cache tools provide programmatic control over the Agent Cache functionality, allowing you to manage cached responses for model-provider endpoint sessions directly from your coding assistant.

Rules

Reverse Proxy

MCP Output Sanitization

The Fiddler Everywhere application provides a sanitization tool that automatically removes sensitive data from the output passed to the MCP server. The data sanitization feature enables users to minimize the risk of disclosing specific information.

Even with the feature enabled, Fiddler will attempt to sanitize the captured traffic, but complete removal of sensitive data is not guaranteed. The user is responsible for verifying the output and preventing unintended disclosure.

By default, the sanitization is enabled for all data passing through the MCP server. You can control what type of data to sanitize or add more strict rules through the Settings > Sanitization section.

MCP Access Policies

The Fiddler Everywhere application provides managed application policies through its Fiddler Everywhere Enterprise tier. The DisableMCP policy enables license administrators to completely restrict access to the Fiddler Everywhere MCP server for licensed users.

Windows

IT teams managing Windows systems can apply app configuration keys using their preferred administrative tooling by setting values in the following registry path:

HKEY_CURRENT_USER\SOFTWARE\Policies\Progress\Fiddler Everywhere

macOS

IT teams managing macOS systems can apply app configuration using their preferred device management solution (such as Jamf, Intune, or similar) by setting the following keys:

For more information on using managed application configurations, refer to the Managed Application Policies article.

See Also

• Fiddler Agent Skills
• Fiddler Prompt Library
• Creating Custom Skills