Quick Start
Deploy your first app with Ink in 2 minutes — connect your AI agent and ship to production
Ink is the first infrastructure platform built for AI agents. Connect your agent, give it a prompt, and get a live URL in minutes.
Ink works however your agent works — through an MCP server (tool-calling), an agent Skill (prompt-guided), or the CLI (command-line). Pick the path that fits your workflow.
1. Create an account
Sign up at ml.ink — then Get Started. The onboarding flow will guide you through the steps below.
2. Install the CLI
The Ink CLI is used by the Skill and is also available as a standalone tool for scripts and CI/CD.
npm install -g @mldotink/clibrew install mldotink/tap/inkThen log in:
ink loginThis opens your browser for authentication. Run ink whoami to verify.
3. Connect your agent
The Ink Skill teaches your agent the full Ink API through a prompt file — no MCP server needed. It uses the CLI under the hood, so there's zero extra infrastructure.
Install the Skill:
npx skills add mldotink/ink-skillThat's it. Your agent now has access to every Ink capability. Use the /ink:ink slash command to trigger the deployment workflow, or just ask your agent to deploy.
Supported clients: Claude Code, OpenClaw, and other AgentSkills-compatible clients.
The MCP server exposes Ink as a set of tools your agent calls via the Model Context Protocol. Works with any MCP-compatible client.
Auth options:
- OAuth — your agent authenticates through a browser flow. No keys to manage.
- Token — create an API key at ml.ink Settings and pass it directly.
Choose your client and auth method below.
claude mcp add --transport http ink "https://mcp.ml.ink/"After installation, start Claude Code and run /mcp to authenticate.
Supported clients: Claude Code, Cursor, Windsurf, VS Code, Claude Desktop, Cline, Codex, Gemini CLI, and any MCP-compatible client.
4. Deploy
Ask your agent to build and deploy something:
Build a React app that shows the current time and deploy it to Ink
Your agent handles everything — writing code, creating a git repo, building the container, and deploying. You'll get a live URL at your-app.ml.ink in seconds.