Best Claude Code MCP Servers in 2026 (Ranked)

The best Claude Code MCP servers in 2026 are the ones that connect Claude to systems it does not already understand natively: GitHub, Linear, Slack, your database, your browser, and your production tooling. This guide ranks the MCP servers that actually expand Claude Code, and it intentionally leaves out things that Claude Code already ships on its own, like built-in file tools and built-in project memory.

Best Claude Code MCP Servers: Top Picks

Server	What it unlocks	Difficulty	When to install
GitHub	Issues, PRs, code search across repos	Easy	Day one
Linear	Tickets, project status	Easy	If your team uses Linear
Slack	Team comms, status updates	Easy	If your team coordinates in Slack
Postgres	Database queries, schema reading	Medium	When the project touches a database
Playwright	Browser automation, E2E testing	Medium	For frontend work
Sentry	Error context, issue triage	Easy	If you ship to production
Context7	Live library docs	Easy	For multi-stack work
Nimbalyst Tracker	In-app planning and bug tracking	Easy	If you use Nimbalyst

The full list of useful Claude Code MCP servers runs to dozens. The eight above cover most working developers.

Two things I would not put in the default list

Two entries show up in a lot of “best MCP server” lists and usually do not belong there for Claude Code:

1. Filesystem MCP server. Claude Code already has built-in file tools (Read, Edit, Write, Glob, Grep) plus fine-grained permission rules and sandboxing. If your goal is “keep Claude away from secrets,” use permissions.deny, sandboxing, or a worktree. An extra filesystem MCP server is only worth it when you specifically want to expose additional directories through MCP or share the same server setup across multiple clients.
2. Memory MCP server. Claude Code already has two built-in memory systems: CLAUDE.md for persistent instructions and auto memory for learned project notes. A separate memory server can still make sense if you want a shared memory layer across different agents or clients, but it is not a day-one Claude Code install.

How to evaluate an MCP server

Before installing anything, three questions:

1. Does it expose tools the agent will actually use? Tools the agent ignores still count toward its tool budget and slow things down. Five well-chosen servers beat twenty.
2. What credentials does it need, and what scope? A server that needs admin tokens in your production database is a server you have to operate carefully. Use read-only roles wherever possible.
3. Is it actively maintained? The MCP standard is young. Servers from January 2025 may not work with Claude Code’s current expectations. Prefer servers with commits in the last three months.

The detailed list

1. GitHub MCP server

The most-installed MCP server, and rightly so. It exposes Claude Code to issues, pull requests, code search, and repository metadata across any repo your token can see.

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

Use a fine-grained personal access token with repo, read:org, and workflow scopes. For team projects, prefer a service account so you can audit Claude Code’s actions separately from your own.

What it unlocks: “Find the issues tagged blocking. Write a PR that closes the top one. Open it as a draft, link the issue, and request review from the owner of the affected module.”

2. Linear MCP server

For teams that live in Linear, this turns Claude Code into a participant in the planning system. It can read tickets, update status, leave comments, and create new issues.

"linear": {
  "command": "npx",
  "args": ["-y", "@linear/mcp-server"],
  "env": { "LINEAR_API_KEY": "lin_api_..." }
}

What it unlocks: “Read all open bugs assigned to me. Pick the one with the smallest scope. Open a fix branch, write the patch, and link the Linear ticket from the PR.”

3. Slack MCP server

For teams that coordinate in Slack. Claude Code can read channel history, post messages, and reply in threads.

"slack": {
  "command": "npx",
  "args": ["-y", "@modelcontextprotocol/server-slack"],
  "env": {
    "SLACK_BOT_TOKEN": "xoxb-...",
    "SLACK_TEAM_ID": "T0123..."
  }
}

What it unlocks: Async status updates. The agent can post when a long-running task finishes, and you can reply with follow-up instructions without leaving Slack.

4. Postgres MCP server

For database-backed projects, exposing the database to Claude Code transforms what it can do. Schema reading, query writing, data inspection, and (if you choose) mutations.

"postgres": {
  "command": "npx",
  "args": [
    "-y",
    "@modelcontextprotocol/server-postgres",
    "postgresql://readonly@localhost/mydb"
  ]
}

Strong recommendation: use a read-only role. Postgres MCP with a read-only DSN is safe and fast. Postgres MCP with a write-capable DSN is a footgun unless you trust the agent fully and have backups.

What it unlocks: “Read the schema. Find the slowest query in the slow_log table. Write a candidate index. Run EXPLAIN ANALYZE before and after.”

5. Playwright MCP server

For frontend and E2E testing work, Playwright MCP lets Claude Code drive a real browser. It can navigate, fill forms, click, screenshot, and assert.

"playwright": {
  "command": "npx",
  "args": ["-y", "@executeautomation/playwright-mcp-server"]
}

What it unlocks: The agent can verify its own UI changes by actually using the app. “Add the new payment form. Then open the checkout page in the browser, fill it out, and confirm the success state appears.”

6. Sentry MCP server

If you ship to production, Sentry MCP wires error context into Claude Code. It can read recent issues, group by frequency, and pull stack traces.

"sentry": {
  "command": "npx",
  "args": ["-y", "@sentry/mcp-server"],
  "env": {
    "SENTRY_AUTH_TOKEN": "sntrys_...",
    "SENTRY_ORG": "my-org"
  }
}

What it unlocks: “Look at this morning’s top three Sentry errors. Find the underlying bug in the code. Open a fix PR for the worst one.”

7. Context7 MCP server

Context7 keeps live, version-aware documentation for hundreds of libraries. The agent can fetch the right docs for the right version on demand.

"context7": {
  "command": "npx",
  "args": ["-y", "@upstash/context7-mcp"]
}

What it unlocks: Less hallucination on library APIs. Instead of guessing the current React or Prisma signature, the agent looks it up.

8. Nimbalyst Tracker MCP server

If you use Nimbalyst, the built-in tracker MCP server gives Claude Code direct access to your in-app plans, tasks, and bugs. The agent can create, update, and close tracker items as part of normal work, with the changes appearing in real time on your kanban board.

What it unlocks: Plans, tasks, and bugs that stay in sync without the agent ever leaving the workspace.

Honorable mentions

• Brave Search for web research without leaving the agent.
• Notion for teams whose source-of-truth lives there instead of Linear.
• Stripe for SaaS work that touches billing.
• Cloudflare for edge and DNS workflows.
• AWS S3 for asset and storage operations.
• Figma for design-to-code handoffs.

The MCP ecosystem doubles every few months. Bookmark the official MCP servers repository for the canonical list.

A sensible starter pack

If you want one configuration to install on day one and tune later:

{
  "mcpServers": {
    "github": { ... },
    "context7": { ... },
    "linear": { ... }
  }
}

Add Slack if your team lives there. Add Postgres, Sentry, or Playwright when the project demands them. Resist installing servers you do not have a clear use for. Each one expands the agent’s tool list, and a bloated tool list hurts the agent’s decision quality.

Claude Code MCP servers in Nimbalyst

Nimbalyst is the open-source visual workspace for building with Codex, Claude Code, and more. The most useful extra MCP server to pair with Nimbalyst is usually GitHub, because Nimbalyst already gives you the visual session layer, tracker, and review workflow. The next-best adds are whatever connect Claude to systems outside the workspace: Linear for tickets, Postgres or Sentry for backend work, and Playwright for UI verification. The MCP configuration you keep in ~/.claude/settings.json still carries over into Nimbalyst sessions, so once you wire those up they show up in both places.

Frequently Asked Questions

What is the best Claude Code MCP server overall?

The GitHub MCP server is the single highest-impact install for most developers. It turns Claude Code from a code generator into a participant in the issue and PR workflow. If you use Nimbalyst, GitHub is also the cleanest complement because Nimbalyst already covers the visual session-management layer.

How many MCP servers should I have configured?

Three to six is the sweet spot for most working developers. Day-one essentials are usually GitHub plus one or two project-specific servers, then Context7 if you frequently work across unfamiliar stacks. More than ten servers tends to slow the agent down without proportional benefit.

Are there official Anthropic MCP servers?

The MCP ecosystem publishes reference servers under the @modelcontextprotocol GitHub org, and many ecosystem servers are built and maintained by the relevant company (Linear, Sentry, Stripe, Cloudflare, and others). Anthropic’s Claude Code docs point to this broader MCP ecosystem rather than limiting you to Anthropic-only extensions.

Can I run Claude Code MCP servers on Windows?

Yes. Most MCP servers ship as npm packages that work on macOS, Linux, and Windows. The configuration syntax is identical. Use absolute paths (no ~ expansion) and forward slashes in the JSON to avoid escaping headaches.

Should I use project-scope or user-scope MCP configs?

User scope (~/.claude/settings.json) for stable servers like GitHub or Context7. Project scope (.claude/settings.json in the repo) for servers that are specific to one codebase, like a database with that project’s DSN. Project scope overrides user scope when both define the same server.

How do I keep MCP secrets out of git?

Two ways. First, put secrets in user-scope settings, which lives in your home directory and is never in a repo. Second, for project scope, commit a settings.example.json with placeholders and add .claude/settings.json to .gitignore. Document the required env vars in the project README.

Related Reading

• Claude Code MCP Setup Guide
• Best MCP Clients in 2026
• Claude Code Subagents Guide
• Claude Code Skills Guide