Skip to main content

Integrate the MCP Server

SF Veritas includes an MCP (Model Context Protocol) server that enables AI coding assistants to query your application's telemetry data. This allows AI tools to help debug issues, analyze performance, and understand your application's behavior using real execution data.

What is MCP?

The Model Context Protocol is an open standard that enables AI applications to connect to external data sources and tools. The SF Veritas MCP server exposes your local telemetry data to AI coding assistants, allowing them to:

  • Query function execution traces and performance metrics
  • Analyze console logs and exceptions
  • Inspect HTTP requests and responses
  • Identify performance bottlenecks and N+1 patterns
  • Debug errors with full stack traces and context

Prerequisites

Before configuring an AI coding assistant, ensure you have:

  1. SF Veritas Desktop App or IDE Extension installed - The MCP server is bundled with both
  2. Telemetry data available - Run your instrumented application at least once to generate data
    • Data is stored in ~/.sailfish/telemetry.db
  3. Node.js 18+ - Required to run the MCP server

MCP Server Location

The MCP server is bundled with both the Desktop App and the VS Code extension.

Desktop App

PlatformPath
macOS/Applications/SF Veritas.app/Contents/Resources/app.asar.unpacked/dist/mcp-server/index.js
Windows%LOCALAPPDATA%\Programs\SF Veritas\resources\app.asar.unpacked\dist\mcp-server\index.js
Linux (.deb)/opt/SF Veritas/resources/app.asar.unpacked/dist/mcp-server/index.js

VS Code Extension

~/.vscode/extensions/sailfish-ai.sf-veritas-extension-*/dist/mcp-server-bundled.js

The * is a glob pattern that matches any version number.

Quick Reference

AI ToolConfig LocationProject-LevelGlobalFormat
Claude Code.mcp.jsonYesYes (~/.claude.json)JSON
OpenAI Codex~/.codex/config.tomlNoYesTOML
Cursor.cursor/mcp.jsonYesYes (~/.cursor/mcp.json)JSON
VS Code + GitHub Copilot.vscode/mcp.jsonYesNoJSON

Configuration Guides