# Vercel MCP

> Vercel MCP brings your continuous deployment infrastructure into chat. Use this MCP to manage projects, domains, and builds directly through your AI agent. You can list existing project architectures, trigger manual deployments from GitHub references, check live build statuses, or map out custom domain DNS settings without leaving your IDE.

## Overview
- **Category:** loved-by-devs
- **Price:** Free
- **Tags:** serverless, deployment, git-integration, infrastructure-as-code, web-hosting, ci-cd

## Description

This connector lets you run complex DevOps commands using only conversation. Forget clicking through multiple dashboards to manage your web applications. Instead, you talk to your agent, and it handles the heavy lifting with Vercel's backend tools.

You can ask for a list of every project in your account or check the status of a build that started hours ago. Need to update a domain alias? Just ask. Or maybe a critical compilation stalled out? Your agent can cancel it immediately, saving you time and compute credits. The whole process is exposed through Vinkius, making deep DevOps control available right where you're coding.

It's about treating your entire CI/CD pipeline like another API call—something conversational and immediate.

## Tools

### cancel_active_build
Stops an ongoing compilation pipeline immediately if something goes wrong.

### trigger_github_deployment
Forces a new Vercel build using a specified GitHub repository reference.

### create_project
Sets up and creates an entirely new Vercel project boundary.

### list_account_domains
Lists all high-level custom domains that are managed by the account.

### list_projects
Provides a list of every Vercel project currently set up in your account.

### delete_project
Permanently removes an entire Vercel project from the platform (irreversible).

### get_deployment_details
Retrieves detailed status and configuration information for a specific build execution.

### get_project_details
Fetches the full configuration details for a specified Vercel project.

### list_project_aliases
Maps out all specific subdomain routing connections for a single project.

### list_deployments
Retrieves the history and status of recent continuous integration builds for a project.

## Prompt Examples

**Prompt:** 
```
List all root domains connected to my Vercel infrastructure.
```

**Response:** 
```
I've queried the apex domains across your environment. You have 3 custom domains actively bound: `production-agent.com`, `internal-tools.io`, and `legacy-blog-marketing.net`. All three signify successful certificate validations.
```

**Prompt:** 
```
Create a manual deploy on the 'billing-service' project pulling directly from the 'main' branch on GitHub repo '341xyz'.
```

**Response:** 
```
I've successfully triggered a queued manual compilation for 'billing-service' using your target GitHub parameters. 

Vercel has acknowledged the deployment intent. The new processing ID assigned to this task is `dpl_H3vM4Q...`. You can poll `get_deployment_details` intermittently with this string or I can check again for you in 3 minutes if you'd prefer.
```

**Prompt:** 
```
Check the status of deployment 'dpl_827a' and give me its exact live preview URL if ready.
```

**Response:** 
```
I've checked deployment `dpl_827a`. Good news – Vercel marks its `readyState` as fully `READY`. There were zero pipeline errors.

You can access your newly compiled live preview here: [https://billing-service-pr-12-test.vercel.app](https://billing-service-pr-12-test.vercel.app)
```

## Capabilities

### Manage project structure
List all existing projects or create brand new Vercel boundaries using simple commands.

### Audit domain records
Map out high-level custom domains and check specific subdomain routing aliases for any given project.

### Control deployments
Check the status of recent builds, retrieve detailed deployment history, or instantly cancel a failing compilation.

### Force builds on demand
Trigger an immediate manual build using specific GitHub repository tags when you need a hotfix fast.

## Use Cases

### Needing to verify a production domain setup
A DevOps Architect needs to confirm if their new staging environment's custom root domain is correctly mapped. Instead of visiting the DNS settings page, they simply ask the agent to use `list_account_domains` and then run `list_project_aliases` for confirmation.

### Catching a broken build in progress
A Frontend Engineer notices the deployment pipeline is stuck. Rather than waiting 30 minutes, they immediately ask the agent to use `cancel_active_build`, stopping the waste of resources and allowing them to fix the source code instead.

### Checking a QA staging link
A Quality Assurance Tester needs the latest URL for review. They prompt the agent, which uses `list_deployments` and pulls the most recent preview URL, saving them from navigating multiple internal links to find the test site.

### Setting up a brand new microservice
A developer needs to start a totally new service. They ask the agent to use `create_project`, supplying the name and framework, which initializes the boundary so they can begin coding immediately after.

## Benefits

- Skip context switching. You don't have to leave your coding environment to manage deployments; simply ask the agent to check status or list projects.
- Control builds instantly. Need to stop a failing compilation? Use the `cancel_active_build` tool to abort it right from the chat interface, saving time and compute.
- Audit domains easily. Instead of digging through multiple dashboards, use the MCP to map out all custom root domains or check specific subdomain routing aliases with one prompt.
- Force deployments when needed. When you're hot-fixing a critical bug, `trigger_github_deployment` lets you bypass standard Git pushes and start a build instantly.
- Get full project status at a glance. You can use `list_deployments` or `get_deployment_details` to see the history and exact current state of any application.

## How It Works

The bottom line is that it turns complex, multi-step web UI interactions into single, conversational prompts.

1. First, subscribe your workspace to the Vercel connector and provide your personal API token for authorization.
2. Next, prompt your AI client with a specific task, such as 'List all projects' or 'Check deployment dpl_xyz'.
3. Finally, your agent runs the necessary commands through this MCP, giving you real-time status updates in the chat.

## Frequently Asked Questions

**How do I use Vercel MCP to list all my projects?**
You ask your agent to run `list_projects`. It will immediately return a comprehensive list of every project boundary you have set up in your Vercel account.

**Can I check the status of an old build using Vercel MCP?**
Yes. You can use `list_deployments` to see recent history, or if you know the ID, ask for details with `get_deployment_details`. It gives a full report on the execution state.

**Does Vercel MCP let me change my domain names?**
It helps you audit them first. You can use `list_account_domains` to see what's attached, and then check specific mappings using `list_project_aliases` before making any changes.

**If my build fails, how do I stop it with Vercel MCP?**
You simply tell the agent to use `cancel_active_build`. It sends the command and stops the resource usage immediately in chat.

**What is the difference between creating a project and listing one using Vercel MCP?**
Using `list_projects` just reads the names of existing projects. Using `create_project` writes data, setting up an entirely new Vercel container for your code.