MCP Integration

Glance implements the Model Context Protocol (MCP), allowing AI agents to interact with network traffic programmatically. This enables powerful AI-assisted debugging and testing workflows.

What is MCP?

The Model Context Protocol is an open standard that allows AI agents to access external tools, resources, and context. Glance's MCP server exposes network traffic data and proxy controls to AI agents like Claude Desktop.

Quick Setup (Claude Code CLI)

Run this one-liner in your terminal to instantly connect Glance to Claude Code:

claude mcp add --transport http glance http://localhost:15502/mcp

Setting Up with Claude Desktop

1. Locate Your Claude Desktop Config

The MCP configuration file is located at:

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

2. Add Glance to MCP Servers

Edit the config file and add Glance as an HTTP/SSE server:

{
  "mcpServers": {
    "glance": {
      "type": "http",
      "url": "http://localhost:15502/mcp"
    }
  }
}

3. Restart Claude Desktop

Restart Claude Desktop to load the new MCP configuration. You should now see Glance listed as an available MCP server.

Setting Up with VS Code (Cline / Roo Code)

1. Open MCP Settings in the extension.
2. Add a new server with type SSE.
3. Paste the URL: http://localhost:15502/mcp

Using Glance with Claude

Once configured, you can ask Claude to interact with your network traffic:

Example Prompts

Inspect recent traffic:

Show me the last 10 HTTP requests captured by Glance

Debug API errors:

Analyze the recent traffic for any failed API calls and help me understand what went wrong

Generate API documentation:

Generate OpenAPI documentation from the captured traffic to the /api/users endpoint

Create mocks:

Create a mock for the /api/login endpoint that returns a success response

Execute requests:

Make a GET request to https://api.github.com/users through the Glance proxy

MCP Capabilities

Resources

Resources provide read-only access to proxy state and traffic data:

Resource	Description
proxy://status	Current proxy address, status, and active rules count
traffic://latest	Most recent 10 HTTP requests in JSON format

Tools

Tools allow AI agents to perform actions:

Tool	Description
inspect_network_traffic	List captured HTTP traffic with filtering and configurable limit
inspect_request_details	Get full headers and body for a specific request
execute_request	Execute custom HTTP requests through the proxy
add_mock_rule	Create a mocking rule to return static responses
add_breakpoint_rule	Create a breakpoint rule to pause traffic
list_rules	List all active mocks and breakpoints
delete_rule	Remove an interception rule by ID
list_scenarios	List all recorded traffic scenarios
get_scenario	Get full sequence details and variable mappings
add_scenario	Create a new scenario with basic metadata
update_scenario	Modify an existing scenario
delete_scenario	Remove a scenario by ID
clear_traffic	Reset/clear captured traffic logs
get_proxy_status	Get real-time proxy address and status

Prompts

Pre-defined prompts for common workflows:

Prompt	Description
analyze-traffic	Analyze recent traffic for errors or anomalies
generate-api-docs	Generate OpenAPI documentation from captured traffic
generate-scenario-test	Transform a recorded scenario into an automated test script

Advanced Usage

Custom MCP Configuration

You can customize Glance's MCP server behavior with additional arguments:

{
  "mcpServers": {
    "glance": {
      "command": "/path/to/glance",
      "args": [
        "--mcp",
        "--mcp-port", "15502",
        "--log-level", "debug"
      ]
    }
  }
}

Using with Other MCP Clients

Glance's MCP server can be used with any MCP-compatible client, not just Claude Desktop. The server runs on http://localhost:15502 by default.

Example Workflows

1. Debug a Failing API Call

1. Run your application with Glance proxy configured
2. Reproduce the failing API call
3. Ask Claude: "Show me the last request to /api/endpoint and help me debug why it's failing"
4. Claude will use inspect_network_traffic and inspect_request_details to analyze the issue

2. Generate Test Code from Scenarios

1. Record a scenario in the Glance dashboard
2. Ask Claude: "Generate a Playwright test script from the 'user-login' scenario"
3. Claude will use the generate-scenario-test prompt to create test code

3. Create Mocks for Testing

1. Capture the real API response in Glance
2. Ask Claude: "Create a mock for this endpoint that returns a similar response but with test data"
3. Claude will use add_mock_rule to create the mock

Troubleshooting

Claude Can't See Glance

1. Verify the path to the Glance binary is correct in your config
2. Check that you restarted Claude Desktop after editing the config
3. Look for error messages in Claude Desktop's MCP logs

MCP Server Won't Start

1. Check if another service is using port 15502
2. Try specifying a different port with --mcp-port
3. Check Glance logs for error messages

Tools Not Working

1. Make sure Glance is running and the proxy is active
2. Verify that traffic is being captured in the dashboard
3. Try using the tools directly in Claude to see specific error messages

Learn More

• MCP Tools & Resources Reference - Detailed API documentation
• Model Context Protocol Documentation - Official MCP docs