---
title: Claude Code setup
description: Configure Claude Code with the Infistar origin and authentication token.
lastUpdated: 2026-09-08
---

Claude Code can read projects, modify files, and run commands. Follow the [official installation instructions](https://code.claude.com/docs/en/setup). The npm installation method is:

```bash
npm install -g @anthropic-ai/claude-code
```

## Configure and start

PowerShell:

```powershell
$env:ANTHROPIC_BASE_URL = "https://infistar.ai"
$env:ANTHROPIC_AUTH_TOKEN = "YOUR_API_KEY"
claude
```

macOS or Linux:

```bash
export ANTHROPIC_BASE_URL="https://infistar.ai"
export ANTHROPIC_AUTH_TOKEN="YOUR_API_KEY"
claude
```

Do not append `/v1`; Claude Code adds the request path.

## Verify

Use `/status` to inspect the connection configuration. Then ask:

```text
Reply only with "Configuration successful". Do not read or modify files, and do not run commands.
```

Ensure the selected model is available to your key and supports the [Messages API](/en/api-reference/claude-format-api).

## Save settings

After verifying the connection, you may store the configuration in `%USERPROFILE%\.claude\settings.json` or `~/.claude/settings.json`:

```json
{
  "env": {
    "ANTHROPIC_BASE_URL": "https://infistar.ai",
    "ANTHROPIC_AUTH_TOKEN": "YOUR_API_KEY"
  }
}
```

This file contains a key. Do not share it or commit it to Git.

For `404`, remove an extra `/v1`. For `401`, check the key and `ANTHROPIC_AUTH_TOKEN`. If the client still requests a login, launch it from the terminal containing these variables and inspect `/status` again. For model errors, confirm the exact model's availability to your key.

See the [official gateway documentation](https://code.claude.com/docs/en/llm-gateway) for client-side configuration details.