How to Set Up MCP (Model Context Protocol): Complete Beginner's Guide
Model Context Protocol (MCP) connects Claude to local files, databases, and APIs. Follow this step-by-step guide to configure MCP on your machine.
In this article
What Is Model Context Protocol?
Large Language Models are normally isolated from the real world. They cannot read your local files, query your database, or check your GitHub issues without custom integrations. MCP solves this problem by providing a standard protocol for clients (like Claude Desktop or Claude Code) to communicate with servers (like a local database or filesystem connector). This means any MCP-compliant server can instantly be used by any MCP-compliant client. Anthropic built MCP with David Soria Parra and Justin Spahr-Summers as the primary architects. The protocol has three core components: a specification and SDK set published on GitHub, local MCP server support built into Claude Desktop, and an open-source repository of pre-built connectors for the most common enterprise tools.
What Comes Pre-Built
Seven enterprise-ready server connectors are available as open-source code you can install immediately. No custom integration work is required to connect Claude to any of them:
- Google Drive: Read, search, and retrieve documents and spreadsheets from your Drive.
- Slack: Access workspace messages, channels, and conversation history.
- GitHub: Retrieve pull requests, issues, and repository content.
- Git: Navigate local git repositories and version history.
- Postgres: Query relational databases and retrieve records directly.
- Puppeteer: Control a headless browser for scraping and automated testing.
- Repository support: Connect to general-purpose code repositories for source navigation.
Who Is Already Using MCP
At launch, Block and Apollo integrated MCP directly into their production systems. Developer platforms including Zed, Replit, Codeium, and Sourcegraph adopted it to enhance their AI-assisted coding environments. Anthropic noted at launch that Claude 3.5 Sonnet is particularly effective at building new MCP server implementations quickly, which helps engineering teams create custom connectors for internal tools and proprietary data sources without starting from scratch. The pre-built server code and full specification are published at modelcontextprotocol.io.
Step 1: Install Claude Desktop or Claude Code
To get started with local MCP, you need either the Claude Desktop application or Claude Code, Anthropic's agentic coding tool. Claude Desktop is the simpler starting point for non-developers and team members who do not work in a terminal. Claude Code, which runs in your terminal or IDE, also supports MCP and is better suited for engineering workflows. While the Claude web interface supports some cloud integrations, local development tools and file access require a desktop or CLI installation to run the local server processes.
To install Claude Code, visit claude.ai and follow the setup instructions for your operating system. Claude Code supports macOS, Windows, Linux, and WSL, with packages available through Homebrew, WinGet, and native Linux package managers. Once installed, navigate to any project directory and type claude to start a session.
Step 2: Locate the Configuration File
Claude Desktop reads its MCP configuration from a JSON file. Depending on your operating system, you can find this file at:
- macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
- Windows: %APPDATA%\Claude\claude_desktop_config.json
Step 3: Configure Your First MCP Servers
Open the claude_desktop_config.json file in a text editor. You can add local servers under the mcpServers key. For example, to add a local filesystem server and a sqlite database server, your config should look like this:
{
"mcpServers": {
"filesystem": {
"command": "npx",
"args": [
"-y",
"@modelcontextprotocol/server-filesystem",
"/Users/username/Projects"
]
},
"sqlite": {
"command": "npx",
"args": [
"-y",
"@modelcontextprotocol/server-sqlite",
"/Users/username/db.sqlite"
]
}
}
}Step 4: Restart Claude Desktop and Verify
After saving the configuration file, restart Claude Desktop. You should see a small plug icon in the prompt input field. Click it to see the list of active MCP servers and the tools they expose. You can now ask Claude to 'list the files in my Projects directory' or 'run a query on my database,' and the model will execute the commands locally via the configured MCP servers.
MCP in Claude Code: Beyond the Desktop
Claude Code uses MCP to connect to external services directly from your terminal or IDE. With MCP configured, Claude Code can read your design documents in Google Drive, update tickets in Jira, pull data from Slack, or connect to your own custom tooling, all within the same session where it reads your codebase and edits files. The Claude Code documentation includes a dedicated MCP quickstart that walks through setting up your first server connection end to end.
The key advantage for engineering teams is that MCP servers follow the same open standard across Claude Desktop and Claude Code. A custom MCP connector you build for an internal service works across both surfaces without additional configuration. Teams building with Claude Code can also use MCP to route tasks from Slack, run automated browser checks via Puppeteer, or pull structured data from a Postgres database, all without leaving their development session.
Availability
MCP support is available across all Claude.ai plans. Any Claude.ai account holder can connect MCP servers to Claude Desktop. Claude for Work customers can run and test local MCP configurations before rolling them out to shared team environments. The full protocol specification, SDK code, and pre-built server connectors are published at modelcontextprotocol.io as open-source projects. SDKs are available for Python, TypeScript, and other languages for teams that need to build custom connectors for internal tools not covered by the seven pre-built options.
Get the Claude playbook in your inbox.
One weekly email for Claude and Claude Code users. Real workflows, no hype. Subscribe and we send you The Claude Power-User Cheatsheet.
— ¶ —
Luke Thompson
Luke Thompson is the founder of The Operations Guide, LLC and editor of The Claude Insider. Based in Jonesborough, Tennessee, he has spent years building AI-augmented business systems and automation workflows for operators and teams. He began working with large language models in production well before the current wave of consumer AI tools, integrating them into client workflows, content pipelines, and operational infrastructure. At The Claude Insider, he writes about Claude with the specificity of someone who uses it daily as a professional tool — not as a reviewer or commentator, but as a builder. His coverage focuses on what actually works: prompt patterns, API integration strategies, agentic workflows, and the real-world tradeoffs that practitioners face. He is not affiliated with Anthropic, PBC.
Articles are researched and drafted with AI assistance, reviewed and edited by Luke Thompson.
Know where AI can pay off in your company.
Take the free two-minute AI Readiness Assessment. See your score, the two gaps holding you back, and the next move worth making.
