consolelog.tools

MCP Server Config Generator

Generate ready-to-paste Model Context Protocol (MCP) server config for Claude Desktop, Claude Code, Cursor, VS Code, Windsurf, and Zed. Add multiple stdio or remote (HTTP/SSE) servers, drop in real published presets (filesystem, GitHub, Postgres, fetch, Puppeteer, Playwright), and the JSON reshapes to each client's exact schema — plus the precise config file path for every OS.

About this ToolHow it works, benefits & use cases

Every MCP-capable app stores Model Context Protocol servers in a slightly different way, and getting one character wrong means the server silently fails to load. This generator removes that guesswork: pick your client — Claude Desktop, Claude Code, Cursor, VS Code, Windsurf, or Zed — and the JSON instantly reshapes to that client’s exact schema. Claude Desktop, Claude Code, Cursor, and Windsurf nest servers under "mcpServers"; VS Code uses "servers" (plus a top-level "inputs" array so secrets are prompted for instead of committed); Zed calls them "context_servers". You can add as many servers as you like, each as either a local stdio command (launched with npx, uvx, node, or any binary, with arguments and environment variables) or a remote HTTP/SSE endpoint with custom headers. A preset gallery drops in real, published servers with correct launch commands — filesystem, GitHub, Postgres, fetch, Puppeteer, and Playwright — fully populated and ready to edit. Alongside the live JSON, the tool shows the precise config file path for the selected client on macOS, Windows, and Linux, so you know exactly where to paste it. Everything runs in your browser, and you can share a link that reproduces your entire setup.

How to Use

  1. 1Pick your target client from the tabs at the top — the output reshapes to that client’s exact JSON schema and shows its config file path.
  2. 2Click "Add server" for a blank entry, or pick a preset (filesystem, GitHub, Postgres, fetch, Puppeteer, Playwright) to drop in a fully populated server.
  3. 3Give the server a name (this becomes its JSON key) and choose a transport: stdio for a local command, or HTTP/SSE for a remote endpoint.
  4. 4For stdio, set the command (e.g. npx) and add argument rows plus any environment variables; for HTTP/SSE, set the URL and add request headers.
  5. 5Read off the config file path shown for your OS, then copy or download the generated JSON and paste it into that file.
  6. 6Restart or reload the client so it picks up the new servers, and use the share link to hand the whole config to a teammate.

Key Benefits

  • One editor, six clients — switch between Claude Desktop, Claude Code, Cursor, VS Code, Windsurf, and Zed instantly
  • Emits each client’s exact root key: mcpServers, servers (+ inputs), or context_servers
  • Supports both local stdio commands and remote HTTP/SSE servers with headers
  • Add unlimited servers with per-server arguments and environment variables
  • Real, published presets with correct npx/uvx commands — no hunting for package names
  • Shows the precise config file location for macOS, Windows, and Linux
  • Runs entirely in the browser with a shareable URL that restores your full setup

Common Use Cases

  • Onboarding a teammate by sharing a link that reproduces your whole MCP setup
  • Migrating the same servers from Claude Desktop to Cursor or VS Code without rewriting JSON by hand
  • Wiring up the filesystem or GitHub server with the correct npx command and token env var
  • Connecting a remote HTTP/SSE MCP server with an Authorization header
  • Generating a project-scoped .mcp.json for Claude Code to commit alongside your repo

Config file · claude_desktop_config.json

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

Edit via Settings → Developer → Edit Config, or open the file directly. Fully quit and reopen Claude Desktop for changes to take effect.

Servers nested under "mcpServers" — paste into your config file

{
  "mcpServers": {
    "filesystem": {
      "command": "npx",
      "args": [
        "-y",
        "@modelcontextprotocol/server-filesystem",
        "/path/to/allowed/dir"
      ]
    },
    "github": {
      "command": "npx",
      "args": [
        "-y",
        "@modelcontextprotocol/server-github"
      ],
      "env": {
        "GITHUB_PERSONAL_ACCESS_TOKEN": "ghp_your_token_here"
      }
    }
  }
}

Servers (2)

Used as the JSON key under the client's servers object

Arguments (3)
0
1
2
Environment variables (0)

No env vars. Add secrets/tokens the server needs at runtime.

Used as the JSON key under the client's servers object

Arguments (2)
0
1
Environment variables (1)

Presets

Real published servers with correct launch commands — adds a fully populated row.

What is MCP?

The Model Context Protocolis an open standard that lets AI apps connect to external tools and data through small "servers". A stdio server is a local command the client spawns (usually via npx or uvx); HTTP/SSE servers are remote endpoints reached over the network. Every client stores these the same way conceptually but with a different JSON key and file location — Claude Desktop, Claude Code, Cursor, and Windsurf use mcpServers, VS Code uses servers (with a top-level inputs array for secrets), and Zed uses context_servers. Pick your client above and this tool emits the exact shape it expects.

Was this tool helpful?

Share Your Experience

Help others discover this tool!

Share on Twitter

Related tools