Getting started with Vercel
Deploy your app on Vercel in three steps: install the CLI, add agent support if you use an AI coding agent, and deploy.
Every Vercel workflow starts with the CLI. Install it whether or not you use an AI coding agent. Agents that can run terminal commands use the CLI to deploy, pull environment variables, and manage projects.
- Terminal
pnpm i -g vercel vercel loginFollow the prompts to authenticate with your Vercel account.
Navigate to your project directory and run:
vercelThe CLI detects your framework, builds your project, and deploys it. To deploy to production:
vercel --prod
See the CLI documentation for the full command reference.
If you use Claude Code or Cursor, install the Vercel Plugin. It gives your agent deployment skills, framework best practices, and slash commands like /vercel-plugin:deploy prod and /vercel-plugin:env.
npx plugins add vercel/vercel-pluginThe plugin activates automatically. No configuration needed.
See the Vercel Plugin documentation for the full list of skills, specialist agents, and slash commands.
If you use Cline, Windsurf, GitHub Copilot, or any of the 18+ agents supported by Skills.sh, install Vercel Skills instead. Skills give your agent the same deployment and framework expertise in a format compatible with your tool.
npx skills add vercel-labs/agent-skillsTo install a specific skill:
npx skills add vercel-labs/agent-skills --skill vercel-react-best-practicesSee Agent Skills for the full list.
You can also deploy without the CLI. Go to the New Project page, connect your GitHub, GitLab, or Bitbucket account, select a repo, and click Deploy. Every push to your connected branch triggers a new deployment automatically.
- Fundamental concepts – How requests, builds, and compute work on Vercel
- Set up environment variables
- Add a custom domain
- Explore supported frameworks
- Vercel Functions – Run server-side code on demand
- Connect the Vercel MCP server – Give AI agents direct access to your Vercel account
- Agent resources – Documentation access, skills, and CLI workflows for AI agents
Was this helpful?