Back
Marketplace/Vercel

TL;DR

Vercel is a Deployment MCP server that lets Claude Code, Cursor, Windsurf and any MCP-compatible AI agent deployments, projects, logs, domains. Install in 1 minute with mcpizy install vercel.

▲

Vercel

Verified 7 tools

Deployment

Last updated May 30, 2026 · By MCPizy team

Deployments, projects, logs, domains

Official homepage

Install Vercel

Via MCPizy CLI (recommended):
mcpizy install vercel
Or run directly:
npx -y @vercel/mcp
View on GitHub

Tools exposed

7 tools available
list_projects

List all Vercel projects in the team

get_project

Get a project's details

Inputs

  • projectIdstringrequired
list_deployments

List deployments for a project

Inputs

  • projectIdstringoptional
get_deployment

Get a single deployment with status

Inputs

  • deploymentIdstringrequired
get_deployment_build_logs

Fetch build logs for a deployment

Inputs

  • deploymentIdstringrequired
get_runtime_logs

Fetch runtime logs for a deployment

Inputs

  • deploymentIdstringrequired
deploy_to_vercel

Deploy the current workspace to Vercel

Inputs

  • projectIdstringoptional

Configuration

Works identically across clients. Only the config file path differs.

Claude Code~/.claude.json
{
  "mcpServers": {
    "vercel": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://mcp.vercel.com"
      ]
    }
  }
}
Cursor.cursor/mcp.json
{
  "mcpServers": {
    "vercel": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://mcp.vercel.com"
      ]
    }
  }
}
Windsurf~/.codeium/windsurf/mcp_config.json
{
  "mcpServers": {
    "vercel": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://mcp.vercel.com"
      ]
    }
  }
}

Authentication setup

  1. 1

    Authenticate via OAuth when prompted (no token setup needed)

    Get it

What you can do with Vercel MCP

Paste any of these prompts into Claude Code, Cursor or another MCP-compatible client.

“List my last 10 deployments and their status”

Uses: list_deployments

“Why did the latest deploy of `mcpizy-dashboard` fail?”

Uses: list_deployments, get_deployment_build_logs

“Show me runtime errors from my production deployment”

Uses: get_runtime_logs

“Trigger a new deployment for my current project”

Uses: deploy_to_vercel

Why Vercel MCP matters

Vercel MCP wraps the Vercel API in agent-callable tools: `list_projects`, `get_deployment`, `list_deployments`, `get_runtime_logs`, `get_deployment_build_logs`, `deploy_to_vercel`, `get_project`. Authentication is a Vercel access token scoped to the user's account or a specific team. The official MCP from the Vercel team is one of the most polished — they treat the agent surface as a first-class platform feature alongside the dashboard and CLI.

We use Vercel MCP daily for three things. First, deploy diagnostics: when a preview deploy 500s, the agent fetches `get_runtime_logs` for the failing function, isolates the error, and proposes the fix. Second, build-log triage: a build fails on a TypeScript error — `get_deployment_build_logs` returns the structured log, the agent identifies the offending file/line, and edits the source. Third, post-merge confirmations: after a Linear MCP closes a ticket, the agent waits for the production deploy to finish (`list_deployments` with a status filter) and posts a Slack confirmation. Token cost is moderate; runtime logs for a 5-min window are ~3-5k tokens, build logs for a successful build are ~2k.

Compared to Netlify MCP or Cloudflare MCP, Vercel wins on Next.js-specific primitives (the agent can read framework-specific build output) and on the polish of the deploy-log API. The honest trade-off: Vercel's pricing and per-team scoping mean the MCP can be confusing when the user has personal projects on a hobby plan and team projects on Pro — the agent has to disambiguate the team slug for every call.

Common pitfalls

Personal projects and team projects use different API paths. A token created at the personal level can't list team projects, and vice versa. If `list_projects` returns empty unexpectedly, check which scope the token belongs to.

`get_runtime_logs` is paginated and time-bounded — by default it returns the last 1 hour. The agent has to explicitly pass a `since` parameter to read further back. Logs older than 1 day on hobby plans are dropped entirely.

Edge Functions and Serverless Functions have different log formats. The MCP returns them as a unified stream, but the agent can't always tell which runtime emitted a given line. For complex debugging, ask Vercel to add the function name to its logger format.

Rate limits: 200 deployments per day on Pro plan; the MCP can hit this if the agent triggers redeployments aggressively. Use `redeploy` (which doesn't count) instead of `deploy_to_vercel` for retries.

How Vercel MCP compares

Honest pros/cons against the closest deployment MCP servers.

ServerStrengthsTrade-offs
Netlify MCPBetter for static / Jamstack sites, simpler Functions modelWeaker for Next.js-specific output, smaller serverless quotas
Cloudflare Workers MCPEdge-first, cheapest at scale, no cold startsDifferent mental model — no Node, Workers runtime only
Railway MCP (community)Full-stack with databases, simpler than Vercel for backendSmaller community, no edge network

Works with

Claude Code
Claude Desktop
Cursor
Windsurf
VS Code + Copilot
Any MCP Client

More Deployment MCPs

🌍

4everland

Decentralized deployment to IPFS and blockchain networks. Web3 hosting.

⚡

EdgeOne Pages

Deploy HTML content to Tencent EdgeOne Pages CDN. Global edge hosting.

Alternatives to Vercel

If Vercel doesn't fit your stack, these Deployment MCP servers solve similar problems.

🌍

Netlify

Site deploys, functions, forms

🚂

Railway

Deploy apps to Railway platform

🟢

Render

Spin up services, query databases, debug with metrics and logs on Render cloud platform.

Key Takeaways

  • Vercel exposes 7 tools for deployment workflows in Claude Code, Cursor and Windsurf.
  • Authentication: 1 step (see setup section above).
  • Install in 1 command: mcpizy install vercel — config written to your client automatically.
  • Free and open source (GitHub source linked above) — verified compatible with every MCP client (Claude Code, Claude Desktop, Cursor, Windsurf, VS Code + Copilot).
  • Best use case: "List my last 10 deployments and their status".

Frequently asked questions

What is the Vercel MCP server?

The Vercel MCP server is an Deployment Model Context Protocol server that lets Claude Code, Cursor, Windsurf, VS Code with Copilot, and other MCP-compatible AI agents deployments, projects, logs, domains. It exposes Vercel's capabilities as tools the AI can call directly from your editor or CLI.

How do I install Vercel MCP with Claude Code?

The fastest way is the MCPizy CLI: run `mcpizy install vercel` and MCPizy will add the server to your `.claude.json` automatically. You can also install it manually by adding an entry under `mcpServers` in `.claude.json` with the command `npx -y @vercel/mcp` and restarting Claude Code.

Is Vercel MCP free?

Yes. The Vercel MCP server is free and open source (see the GitHub repository linked on this page). You may still need a Vercel account or API key to connect the server to the underlying service, but the MCP layer itself has no MCPizy subscription cost.

Does Vercel MCP work with Cursor and Windsurf?

Yes. Any MCP-compatible client works — including Claude Code, Claude Desktop, Cursor (via `.cursor/mcp.json`), Windsurf, VS Code with Copilot Chat, and custom agents built on the MCP SDK. The same install command targets all of them; only the config file path differs.

What can I do with Vercel MCP?

Once installed, your AI agent can deployments, projects, logs, domains directly inside your conversation. Typical use cases include asking Claude Code or Cursor to run Vercel operations, inspect results, chain Vercel with other MCP servers (see our Workflow Recipes), and automate repetitive deployment tasks without leaving your editor.

Manage all your MCPs in one place

Monitor usage, track costs, and discover new MCPs.

Get Started Free