Interactive Dev Tool

MCP Config Generator

Build, configure, and validate production-ready JSON configs for Claude Desktop, Cursor, Windsurf, and AI Agents in seconds.

Target Client:

Select MCP Servers (2 active)

GitHubOfficial

Dev Tools

Manage repositories, issues, pull requests, file contents, and search code.

PostgreSQLPopular

Databases

Query and inspect PostgreSQL schemas, tables, views, and execute read-only queries.

SQLiteOfficial

Databases

Query and inspect local SQLite databases, explore table schemas and indexes.

FilesystemEssential

Core

Read, write, search, and manage directories and files with scoped sandbox access.

Brave SearchPopular

Search & Web

Web and local search capability powered by the privacy-preserving Brave Search API.

Memory GraphOfficial

Core

Knowledge-graph persistent memory system to maintain entity relationships across sessions.

Web FetchOfficial

Search & Web

Fetch web pages and convert HTML content to clean, token-efficient Markdown.

Puppeteer (Browser)Official

Search & Web

Full headless Chrome browser automation, DOM traversal, and screenshot capture.

SentryObservability

Dev Tools

Query Sentry error issues, crash events, stack traces, and release telemetry.

SlackCommunication

Dev Tools

Interact with Slack channels, read thread history, and post agent updates.

Sequential ThinkingReasoning

Core

Dynamic multi-step reasoning tool enabling structured problem decomposition.

Docker EngineDevOps

Dev Tools

List, inspect, start, stop containers, and manage Docker volumes and images.

Customize Server Parameters & Secrets

GitHub Settingsnpx
Filesystem Settingsnpx

Custom MCP Servers

Add proprietary, internal, or community servers not in standard catalog.

claude_desktop_config.json2 servers
{
  "mcpServers": {
    "github": {
      "command": "npx",
      "args": [
        "-y",
        "@modelcontextprotocol/server-github"
      ],
      "env": {
        "GITHUB_PERSONAL_ACCESS_TOKEN": "<YOUR_GITHUB_PERSONAL_ACCESS_TOKEN>"
      }
    },
    "filesystem": {
      "command": "npx",
      "args": [
        "-y",
        "@modelcontextprotocol/server-filesystem",
        "C:/Projects"
      ]
    }
  }
}
Installation Locations for Claude Desktop:
windows%APPDATA%\Claude\claude_desktop_config.json
mac~/Library/Application Support/Claude/claude_desktop_config.json
linux~/.config/Claude/claude_desktop_config.json

Paste this JSON directly into your client's config file and restart the application to activate your MCP tools.

Breaking Schema & CVE Alerts

MCP server tools and protocol versions update weekly. Get pinged when breaking JSON-RPC changes or security advisories affect your stack.

Technical Architecture Guide

How MCP Client Configuration Works

The Model Context Protocol establishes a bidirectional JSON-RPC 2.0 communication channel between an AI Host application (such as Claude Desktop, Cursor, or the OpenAI Codex CLI) and standalone MCP server processes. Rather than compiling plugins or hardcoding API connectors, the client dynamically spawns servers as child processes based on your configuration file.

Stdio Child Process Spawning

In local environments, MCP clients execute the binary specified in command (typically npx, node, uvx, or python) and frame JSON-RPC messages across stdin and stdout. The server process inherits only the environment variables declared in your config.

Security Sandboxing & Scopes

Every server operates within an isolated OS process boundary. File servers (such as @modelcontextprotocol/server-filesystem) strictly enforce path restrictions based on arguments passed at launch, preventing models from traversing into sensitive parent directories or system files.

Common Configuration Traps & Troubleshooting

  • 1.GUI $PATH Invisibility (ENOENT Error): Desktop GUI applications on macOS and Windows do not inherit user shell aliases or custom PATH modifications. If Claude crashes with spawn npx ENOENT, replace "npx" with the absolute system binary path (e.g. C:\\Program Files\\nodejs\\npx.cmd or /usr/local/bin/npx).
  • 2.JSON Syntax Errors: Trailing commas after the last server definition or unescaped backslashes in Windows file paths will cause client JSON parsers to abort silently during application boot.
  • 3.Environment Variable Precedence: When deploying across production teams, avoid hardcoding database passwords directly in shared configuration files; use environment variables or secret managers to inject credentials at runtime.

MCP Configuration FAQ

What is an MCP configuration file?

An MCP configuration file (like claude_desktop_config.json or mcp.json) tells your AI client which MCP servers to launch, what command and arguments to use, and any required environment variables such as API keys or connection strings.

Where is the MCP config file located for Claude Desktop?

On Windows: %APPDATA%\Claude\claude_desktop_config.json. On macOS: ~/Library/Application Support/Claude/claude_desktop_config.json. On Linux: ~/.config/Claude/claude_desktop_config.json.

How do I add multiple MCP servers to one config?

In your client configuration file, add additional named keys under the 'mcpServers' object. Each server entry needs its own 'command', 'args', and optional 'env' variables.

Why do I need to restart my AI client after modifying MCP config?

Most desktop AI clients (like Claude Desktop) read the MCP configuration file only on initial application startup. After making edits to the JSON or TOML file, a full restart is required to spawn the new server processes.