Best MCP Servers for Developers: GitHub,

Best MCP Servers for Developers: GitHub, Docker, Kubernetes & More

MCP turns your AI coding assistant into something that actually knows your infrastructure. Instead of describing your setup to Claude or Cursor, MCP servers let the AI read your GitHub repos, manage your containers, query your monitoring tools, and track your issues — directly.

Here are the MCP servers that matter most for developers, organized by what they do.

Source Control & Code Review

GitHub MCP Server (Official)

This is the most-used MCP server in the ecosystem, and for good reason. It gives your AI full access to your GitHub workflow.

What you can do with it:

Create and manage issues and PRs
Read PR diffs and review code
Analyze CI/CD failure logs
Search code across repositories
Manage branches and commits
Triage issues with AI assistance

Install (Docker — recommended for security):

json
{
  "mcpServers": {
    "github": {
      "command": "docker",
      "args": [
        "run", "-i", "--rm",
        "-e", "GITHUBPERSONALACCESSTOKEN=ghpyour_token",
        "ghcr.io/github/github-mcp-server"
      ]
    }
  }
}

Install (npx — simpler):

json
{
  "mcpServers": {
    "github": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-github"],
      "env": {
        "GITHUBPERSONALACCESSTOKEN": "ghpyour_token"
      }
    }
  }
}

Pricing: Free, open source Real use case: Ask your AI "Review the latest PR and summarize what changed" or "Why is CI failing on this branch?" Instead of switching between tabs, the AI reads the diff and logs directly.

See our full GitHub MCP Server review.

GitLab MCP Server

Similar to the GitHub server, but for GitLab. Manage merge requests, issues, pipelines, and repository content.

Install:

json
{
  "mcpServers": {
    "gitlab": {
      "command": "npx",
      "args": ["-y", "gitlab-mcp-server"],
      "env": {
        "GITLAB_URL": "https://gitlab.com",
        "GITLAB_TOKEN": "your-token"
      }
    }
  }
}

Pricing: Free, open source Best for: Teams on GitLab who want the same MCP-powered workflow as GitHub users

Containers & Infrastructure

Docker MCP Server

Manage your entire Docker environment through AI — containers, images, volumes, networks, and compose stacks.

What you can do:

List and inspect running containers
Start/stop/restart containers
View container logs
Manage Docker Compose stacks
Build and push images
Debug container networking

Install: Docker MCP is available through Docker Desktop's built-in MCP support. Enable it in Docker Desktop settings under the MCP section. Pricing: Free (included with Docker Desktop) Real use case: "Why is my API container restarting?" The AI reads the container logs, checks the health status, and tells you what's wrong — all without you opening a terminal.

Kubernetes MCP Server

Interact with Kubernetes and OpenShift clusters. Maintained by Red Hat's containers organization.

What you can do:

Get pod status and logs
Manage deployments, services, and configmaps
Troubleshoot cluster issues
Scale workloads
View resource usage

Install:

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

The server uses your existing ~/.kube/config for authentication.

Pricing: Free, open source Best for: DevOps engineers and platform teams managing K8s clusters

Terraform MCP Server

Read and manage Terraform configurations. Understand state files, plan changes, and navigate infrastructure-as-code.

Install:

json
{
  "mcpServers": {
    "terraform": {
      "command": "npx",
      "args": ["-y", "tfmcp"]
    }
  }
}

Pricing: Free, open source Best for: Infrastructure engineers working with Terraform configurations

Portainer MCP Server

Manage Portainer instances through AI — container management, deployment operations, and infrastructure monitoring across multiple Docker hosts.

Pricing: Free, open source Best for: Teams using Portainer to manage Docker environments across multiple servers

Error Tracking & Monitoring

Sentry MCP Server

Pull error reports, stack traces, and performance data from Sentry into your AI context.

What you can do:

View recent errors and their frequency
Read full stack traces
Analyze error trends
Get AI-powered fix suggestions based on real error data

Install:

json
{
  "mcpServers": {
    "sentry": {
      "command": "npx",
      "args": ["-y", "sentry-mcp-server"],
      "env": {
        "SENTRYAUTHTOKEN": "your-token",
        "SENTRY_ORG": "your-org"
      }
    }
  }
}

Pricing: Free server (Sentry has its own pricing) Real use case: "What's the most common error in production this week?" The AI pulls real Sentry data and gives you a prioritized list with fix suggestions.

Datadog MCP Server

Query Datadog metrics, logs, and APM data through AI.

Best for: Teams using Datadog for infrastructure and application monitoring

Prometheus MCP Server

Query Prometheus metrics and alert rules. Great for understanding system performance trends.

Best for: Teams running Prometheus for metrics collection

Project Management

Linear MCP Server

Manage Linear issues, projects, and sprints through AI.

What you can do:

Create and update issues
View project status and sprint progress
Search issues by label, assignee, or status
Generate sprint summaries

Install:

json
{
  "mcpServers": {
    "linear": {
      "command": "npx",
      "args": ["-y", "linear-mcp-server"],
      "env": {
        "LINEARAPIKEY": "your-key"
      }
    }
  }
}

Pricing: Free server (Linear has its own pricing) Best for: Engineering teams using Linear who want to manage issues without leaving their editor

Jira / Atlassian MCP Server

Atlassian's official Rovo MCP server connects to Jira, Confluence, and other Atlassian products.

Install (Streamable HTTP):

json
{
  "mcpServers": {
    "atlassian": {
      "url": "https://mcp.atlassian.com/v1/mcp",
      "headers": {
        "Authorization": "Bearer your-token"
      }
    }
  }
}

Note: Atlassian is deprecating the old SSE endpoint (/v1/sse) after June 2026. Use the Streamable HTTP endpoint (/v1/mcp) instead.

Pricing: Free (included with Atlassian Rovo) Best for: Teams on Jira/Confluence who want AI-assisted project management

Code Intelligence

Context7 MCP Server

Provides up-to-date documentation for 9,000+ libraries. This solves the #1 problem with AI coding assistants: outdated API suggestions.

Install:

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

Pricing: Free Real use case: When your AI suggests a deprecated React API, Context7 feeds it the current docs so you get working code on the first try.

JetBrains MCP Server

Expose JetBrains IDE features — intelligent refactoring, code navigation, and project configuration — to any MCP client.

Pricing: Free (requires JetBrains IDE license) Best for: JetBrains users who also use Claude Desktop or another MCP client alongside their IDE

BifrostMCP (VS Code)

A VS Code extension that exposes semantic tools like Find Usages, Go to Definition, and Rename to LLM clients via MCP.

Pricing: Free, open source Best for: VS Code users who want AI to understand code structure, not just text

Recommended Developer Stacks

Full-Stack Web Developer:

GitHub + PostgreSQL + Context7 + Sequential Thinking

DevOps / Platform Engineer:

Kubernetes + Docker + Terraform + Prometheus + GitHub

Mobile Developer:

GitHub + Sentry + Linear + Filesystem

Data Engineer:

PostgreSQL + MongoDB + GitHub + Brave Search

Engineering Manager:

Linear + GitHub + Slack + Notion

Installation Tips

Start with 2-3 servers. Too many servers means too many tools for the AI to consider, which slows down responses.

Use Docker for sensitive servers. The GitHub Docker image is more secure than npx because it runs in an isolated container.

Per-project configs. Put database and project-specific servers in .cursor/mcp.json or .vscode/mcp.json at the project level. Keep universal servers (GitHub, Context7) in global config.

Scope your tokens. Create fine-grained personal access tokens with minimal permissions. A GitHub token for MCP doesn't need admin access.

What's Next

See the full MCP servers guide for servers beyond dev tools
Connect your databases: best MCP database servers
Read the MCP security guide before connecting production infrastructure
Compare MCP clients to pick the right editor

We track MCP support across 500+ AI tools in our directory. Browse MCP-compatible tools to find the right servers for your stack.

Best MCP Servers for Developers: GitHub, Docker, Kubernetes & More