> ## Documentation Index
> Fetch the complete documentation index at: https://cli-docs.relai.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Install & setup

> Install the CLI, confirm prerequisites, and configure credentials — three commands and you're ready to initialize a project.

<Tabs sync={false}>
  <Tab title="macOS 15+ (Apple silicon), Linux">
    <Steps>
      <Step title="Install the CLI">
        Install the RELAI CLI with the standalone installer for macOS or Linux.

        ```sh theme={"system"}
        curl -fsSL https://api.relai.ai/install.sh | bash
        ```
      </Step>

      <Step title="Configure credentials">
        Run setup after installing. It checks prerequisites, asks for your RELAI API credentials, and saves your theme in `~/.relai/config.toml`. It also installs plugins for any detected coding agents (Codex or Claude Code).

        <Info>
          The default API URL is `https://api.relai.ai`.
        </Info>

        <Info>
          Create or manage CLI keys from the [API keys page](https://platform.relai.ai/settings/workspace/api-keys). Never paste or commit API keys.
        </Info>

        ```sh theme={"system"}
        relai setup
        ```
      </Step>
    </Steps>
  </Tab>
</Tabs>

## Prerequisites

RELAI supports Python agents, plus TypeScript and Go agents experimentally. `relai setup --check` (see [setup](cli/setup)) audits the requirements for Python and TypeScript agents.

<Tabs sync={false}>
  <Tab title="Universal">
    Required for every project:

    * [`git`](https://git-scm.com/) — your project must be a git repository with at least one commit. [Install](https://git-scm.com/install/)

    * Bash — used by generated simulator scripts.
      Recommended for optimizer pull requests:

    * [`gh`](https://cli.github.com/) — enables automatic optimizer PRs. [Install](https://github.com/cli/cli#installation)
  </Tab>

  <Tab title="Python">
    Required for Python agents:

    * [`python3`](https://www.python.org/) 3.11 or newer. [Install](https://www.python.org/downloads/)
    * [`curl`](https://curl.se/) — downloads the RELAI Python SDK. [Install](https://curl.se/download.html)
    * [`jq`](https://jqlang.org/) — JSON processing. [Install](https://jqlang.org/download/)

    Recommended:

    * [`uv`](https://docs.astral.sh/uv/) — creates an isolated simulator environment without touching your system Python. [Install](https://docs.astral.sh/uv/getting-started/installation/)
  </Tab>

  <Tab title="TypeScript (experimental)">
    Required for TypeScript agents:

    * [Node.js](https://nodejs.org/) 20 or newer.
    * At least one package manager: `npm`, `pnpm`, Yarn, or Bun.
  </Tab>

  <Tab title="Go (experimental)">
    Required for Go agents:

    * [Go](https://go.dev/dl/) 1.24.4 or newer.
  </Tab>
</Tabs>

<Warning>
  RELAI is in beta. For adapter capability requirements, constraints, and known limitations, see [Compatibility](/compatibility).
</Warning>

Installed and configured? Next:

<CardGroup cols={2}>
  <Card title="Quickstart" icon="bolt" href="/quickstart">
    Run the loop on your own agent: init → learning-env → simulate → optimize.
  </Card>

  <Card title="Try the sample agent" icon="flask" href="/sample-agent">
    No project of your own yet? Walk the same loop in your browser.
  </Card>
</CardGroup>