Skip to content

Pre-GA Design Partner and Early Access only. Request access

Helix Docs

Getting started

The path from an empty terminal to a deployed Helix app: what Helix is, the quickstart, building with any coding agent, and project anatomy.

For builders
View as Markdown

The getting-started path takes you from an empty terminal to a live Helix app. It covers the platform, the setup journey, the agent workflow, and the project layout. Work through the pages in order; the path assumes no prior Helix experience. It works on macOS, Linux, and Windows, with Claude Code or any other MCP-capable coding agent.

What you end up with: a deployed app at https://{project-id}.helix-app.ai, managed from the dashboard at app.helix.tray.ai. Deployed apps run behind login, so only people in your Tray or Helix organization who you grant access can open them.

Prerequisites at a glance

You need four things before you start:

The quickstart walks through checking each one.

The path

The quickstart stands alone: if you’d rather build first and read later, jump straight to it.

  1. What is Helix? What the platform does, who it’s for, and what’s included.
  2. Quickstart Install the CLI, log in, scaffold a project, and deploy it.
  3. Building with Claude Code How the agent scaffolds, tests, and deploys through the CLI’s MCP server, which the workflow requires.
  4. Building with other coding agents Point Cursor, VS Code, Codex, or another MCP client at the same project instead.
  5. Project structure What helix init creates and what each file is for.

If you get stuck

Troubleshooting collects the common failure modes: install and PATH problems, Node version issues, port conflicts, and login errors. Each entry gives the cause and the fix commands.

Loading search…

Jump to a section

tab to move · esc to close