Clients & Integrations
Guide

Best MCP Servers for Cursor IDE in 2026

The best MCP servers for Cursor IDE — top 15 servers with .cursor/mcp.json configs, workflow tips, and performance notes for developers.

11 min read
Updated February 26, 2026
By MCPServerSpot Team

The best MCP servers for Cursor are development-focused tools that enhance Cursor's AI Composer and Chat with capabilities like database querying, GitHub management, browser testing, documentation search, and infrastructure control. Unlike general-purpose MCP setups, Cursor users benefit most from servers that integrate directly into coding workflows -- reading project files, running tests, managing deployments, and automating repetitive development tasks.

This article covers the top 15 MCP servers specifically selected for Cursor IDE users. For the full guide on setting up MCP in Cursor and other IDEs, see MCP in Cursor, VS Code and Other IDEs.

How Cursor Uses MCP Servers

Before diving into the server list, it is worth understanding how Cursor integrates MCP differently from Claude Desktop:

FeatureCursor MCPClaude Desktop MCP
Primary modeComposer (multi-file editing) and ChatConversational chat
Context awarenessFull codebase indexing plus MCP toolsMCP tools only (no codebase indexing)
Config location.cursor/mcp.json in project rootclaude_desktop_config.json in system directory
Tool invocationAI decides based on coding contextAI decides based on conversation context
Best forDevelopment workflows, code generation, refactoringGeneral productivity, analysis, writing

Cursor's Composer mode can combine codebase context with MCP tool results, making development-focused servers especially powerful.


Top 15 MCP Servers for Cursor

1. GitHub Server

Best for: Pull requests, code reviews, issue management directly from your editor

The GitHub server is the top pick for Cursor users because it connects your coding workflow directly to your repository. Create PRs from Composer, search issues while debugging, and review diffs without leaving the editor.

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

Key tools: create_pull_request, search_repositories, get_file_contents, create_issue, list_commits

Cursor workflow tip: After making changes in Composer, ask Cursor to create a PR with a description summarizing the changes. The GitHub server handles it without switching to the browser.


2. Filesystem Server

Best for: Reading and writing files outside Cursor's workspace

While Cursor already has codebase access, the Filesystem server extends Claude's reach to directories outside your project -- configuration files, logs, documentation folders, and related repositories.

{
  "mcpServers": {
    "filesystem": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-filesystem", "/Users/you/projects", "/Users/you/configs"]
    }
  }
}

Key tools: read_file, write_file, list_directory, search_files


3. PostgreSQL / SQLite Database Servers

Best for: Querying databases while writing application code

Database servers are essential for full-stack developers using Cursor. When writing API endpoints or data models, having Claude query the actual database schema and data eliminates guesswork.

{
  "mcpServers": {
    "postgres": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-postgres"],
      "env": {
        "POSTGRES_CONNECTION_STRING": "postgresql://user:pass@localhost:5432/mydb"
      }
    }
  }
}

Key tools: query, list_tables, describe_table

Cursor workflow tip: Ask Cursor to check the database schema before generating migration files or ORM models. This ensures the generated code matches your actual database structure.

See more in our Database and Vector DB Servers overview.


4. Brave Search Server

Best for: Looking up documentation, API references, and error solutions

When Cursor's AI encounters an unfamiliar library or error message, the Brave Search server lets it look up current documentation and Stack Overflow solutions in real time.

{
  "mcpServers": {
    "brave-search": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-brave-search"],
      "env": {
        "BRAVE_API_KEY": "your_brave_api_key"
      }
    }
  }
}

Key tools: brave_web_search, brave_local_search


5. Puppeteer Browser Automation Server

Best for: Testing web applications, taking screenshots, scraping pages

Puppeteer gives Cursor a headless browser for automated testing. After generating frontend code, ask Cursor to navigate to your dev server, take a screenshot, and verify the output.

{
  "mcpServers": {
    "puppeteer": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-puppeteer"]
    }
  }
}

Key tools: puppeteer_navigate, puppeteer_screenshot, puppeteer_click, puppeteer_evaluate

Check out Browser Automation Servers for more options.


6. Docker Server

Best for: Managing development containers and Docker Compose stacks

The Docker server lets Cursor manage your containerized development environment -- starting services, checking logs, and troubleshooting container issues without switching to the terminal.

{
  "mcpServers": {
    "docker": {
      "command": "npx",
      "args": ["-y", "docker-mcp-server"]
    }
  }
}

Key tools: list_containers, start_container, stop_container, container_logs, build_image


7. Sentry Server

Best for: Debugging production errors while writing fixes

Sentry integration lets Cursor pull error details, stack traces, and frequency data directly into your coding session. When fixing bugs, Claude can fetch the exact error context before suggesting a patch.

{
  "mcpServers": {
    "sentry": {
      "command": "npx",
      "args": ["-y", "@sentry/mcp-server"],
      "env": {
        "SENTRY_AUTH_TOKEN": "your_sentry_token"
      }
    }
  }
}

Key tools: search_issues, get_issue_details, list_projects

Explore more in Developer Tools Servers.


8. Linear Server

Best for: Managing sprint tasks and issues alongside code changes

Linear's MCP server connects your project management board to Cursor. Create issues for discovered bugs, update task statuses after completing work, and search backlogs for context on what needs building.

{
  "mcpServers": {
    "linear": {
      "command": "npx",
      "args": ["-y", "linear-mcp-server"],
      "env": {
        "LINEAR_API_KEY": "lin_api_your_key"
      }
    }
  }
}

Key tools: create_issue, search_issues, update_issue, list_projects


9. Fetch / HTTP Server

Best for: Calling APIs, fetching documentation pages, testing endpoints

The Fetch server lets Cursor make HTTP requests -- useful for testing API endpoints you are building, fetching remote documentation, or pulling data from external services.

{
  "mcpServers": {
    "fetch": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-fetch"]
    }
  }
}

Key tools: fetch


10. Notion Server

Best for: Accessing project documentation and specs while coding

When your team stores technical specs, API documentation, or architecture decisions in Notion, this server lets Cursor read that context while generating code.

{
  "mcpServers": {
    "notion": {
      "command": "npx",
      "args": ["-y", "notion-mcp-server"],
      "env": {
        "NOTION_API_KEY": "ntn_your_key_here"
      }
    }
  }
}

Key tools: search_pages, read_page, create_page


11. Supabase Server

Best for: Managing Supabase projects, database, auth, and storage

For teams using Supabase as their backend, this server gives Cursor direct access to database operations, auth management, and storage -- all while writing application code.

{
  "mcpServers": {
    "supabase": {
      "command": "npx",
      "args": ["-y", "supabase-mcp-server"],
      "env": {
        "SUPABASE_URL": "https://your-project.supabase.co",
        "SUPABASE_SERVICE_KEY": "your_service_key"
      }
    }
  }
}

Key tools: query, list_tables, manage_auth_users, list_storage_buckets


12. Vercel Server

Best for: Managing deployments, checking build status, reviewing logs

The Vercel MCP server lets Cursor interact with your deployment pipeline -- checking build statuses, reading deployment logs, and managing environment variables.

{
  "mcpServers": {
    "vercel": {
      "command": "npx",
      "args": ["-y", "vercel-mcp-server"],
      "env": {
        "VERCEL_TOKEN": "your_vercel_token"
      }
    }
  }
}

Key tools: list_deployments, get_deployment_logs, list_projects, manage_env_vars


13. Slack Server

Best for: Sharing updates, searching conversations for context

The Slack server keeps your team communication accessible within Cursor. Search for prior discussions about a feature, or post updates when you complete a task.

{
  "mcpServers": {
    "slack": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-slack"],
      "env": {
        "SLACK_BOT_TOKEN": "xoxb-your-token",
        "SLACK_TEAM_ID": "T0123456789"
      }
    }
  }
}

Key tools: send_message, search_messages, list_channels


14. Sequential Thinking Server

Best for: Breaking down complex architectural decisions and refactoring plans

This server enhances Cursor's reasoning on complex tasks by providing structured step-by-step thinking. Useful for architectural planning, complex refactoring, and debugging multi-layered issues.

{
  "mcpServers": {
    "sequential-thinking": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-sequential-thinking"]
    }
  }
}

Key tools: sequentialthinking


15. AWS / Cloudflare Server

Best for: Managing cloud infrastructure from your editor

Cloud infrastructure servers let Cursor interact with your deployment environment -- checking service status, reading logs, and managing resources without leaving the IDE.

{
  "mcpServers": {
    "aws": {
      "command": "npx",
      "args": ["-y", "aws-mcp-server"],
      "env": {
        "AWS_ACCESS_KEY_ID": "your_key",
        "AWS_SECRET_ACCESS_KEY": "your_secret",
        "AWS_REGION": "us-east-1"
      }
    }
  }
}

Key tools: list_resources, describe_instance, get_logs

See Cloud Provider Servers for more options.


Quick Comparison Table

ServerPrimary UseSetupFreeCursor-Specific Value
GitHubVersion controlEasyYesPR creation from Composer
FilesystemFile accessEasyYesCross-project file access
PostgreSQL/SQLiteDatabaseMediumYesSchema-aware code generation
Brave SearchWeb searchEasyYesReal-time documentation lookup
PuppeteerBrowserEasyYesVisual testing of generated UI
DockerContainersEasyYesDev environment management
SentryError trackingEasyYesBug context while writing fixes
LinearTask managementEasyYesIssue tracking in editor
FetchHTTP requestsEasyYesAPI endpoint testing
NotionDocumentationEasyYesSpec access while coding
SupabaseBackendEasyYesFull-stack development
VercelDeploymentEasyYesDeployment management
SlackCommunicationMediumYesTeam context search
Sequential ThinkingReasoningEasyYesComplex architectural planning
AWS/CloudflareCloud infraMediumYesInfrastructure management

Performance Considerations for Cursor

Cursor runs MCP servers as background processes, just like Claude Desktop. However, coding sessions tend to involve more frequent tool calls, so performance matters more:

TipDetails
Start with 3-5 serversAdding all 15 at once increases memory usage and startup time; add servers as you need them
Use project-specific configsDifferent projects need different servers; keep .cursor/mcp.json tailored per repository
Monitor memoryEach Node.js-based server uses 30-80 MB of RAM; Python servers may use more
Restart after config changesCursor must be restarted or the window reloaded after editing .cursor/mcp.json
Share configs via GitCommit .cursor/mcp.json (without secrets) to your repo so teammates get the same server setup

Recommended Starter Configurations

Full-Stack Web Developer

{
  "mcpServers": {
    "github": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-github"],
      "env": {
        "GITHUB_PERSONAL_ACCESS_TOKEN": "ghp_token"
      }
    },
    "postgres": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-postgres"],
      "env": {
        "POSTGRES_CONNECTION_STRING": "postgresql://localhost:5432/mydb"
      }
    },
    "puppeteer": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-puppeteer"]
    },
    "brave-search": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-brave-search"],
      "env": {
        "BRAVE_API_KEY": "your_key"
      }
    }
  }
}

DevOps / Infrastructure Engineer

{
  "mcpServers": {
    "docker": {
      "command": "npx",
      "args": ["-y", "docker-mcp-server"]
    },
    "aws": {
      "command": "npx",
      "args": ["-y", "aws-mcp-server"],
      "env": {
        "AWS_ACCESS_KEY_ID": "your_key",
        "AWS_SECRET_ACCESS_KEY": "your_secret",
        "AWS_REGION": "us-east-1"
      }
    },
    "github": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-github"],
      "env": {
        "GITHUB_PERSONAL_ACCESS_TOKEN": "ghp_token"
      }
    },
    "sentry": {
      "command": "npx",
      "args": ["-y", "@sentry/mcp-server"],
      "env": {
        "SENTRY_AUTH_TOKEN": "your_token"
      }
    }
  }
}

If servers are not connecting or tools are not appearing, see our dedicated Cursor MCP Troubleshooting Guide for step-by-step fixes.

What to Read Next