Using OpenVeer in Claude Code

Prerequisites

Claude Code is a command-line AI programming assistant by Anthropic that supports direct AI conversation, code generation, and debugging in the terminal.
By connecting to OpenVeer, you can use multiple models including GPT, Claude, and Gemini within Claude Code.

Before you begin, make sure you have:

1. Obtained an OpenVeer API Key
   Log in to the OpenVeer Console to get your API key (starts with sk-)

Step 1: Install Claude Code

Choose any of the following methods to install:

macOS / Linux (Recommended)

Install with the official script:

curl -fsSL https://claude.ai/install.sh | bash

Or install via Homebrew:

brew install --cask claude-code

Windows

PowerShell:

irm https://claude.ai/install.ps1 | iex

CMD:

curl -fsSL https://claude.ai/install.cmd -o install.cmd && install.cmd && del install.cmd

NPM Install

If you have Node.js 18 or newer installed, you can install via npm:

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

Works on all operating systems.

Verify Installation

After installation, run the following command to confirm:

claude --version

If a version number is displayed (e.g., 1.x.x), the installation was successful.

Step 2: Configure OpenVeer API

Three configuration methods are available. Choose the one that suits your workflow.

Method 1: Edit settings.json (Recommended)

The most stable approach — configure once and it persists.

1. Locate the configuration directory:

• Windows: Press Win + R, enter %userprofile%\.claude
• macOS: Press Command + Shift + G, enter ~/.claude
• Linux: Navigate to ~/.claude

2. Create or edit the settings.json file:

{
  "env": {
    "ANTHROPIC_BASE_URL": "https://api.openveer.com",
    "ANTHROPIC_AUTH_TOKEN": "sk-xxxxxxxxxxxx",
    "ANTHROPIC_MODEL": "claude-opus-5",
    "CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC": "1"
  }
}

Save the file and restart Claude Code to apply.

Method 2: Permanent Environment Variables

Write configuration to system environment so all terminal windows load it automatically.

macOS (zsh)

echo 'export ANTHROPIC_BASE_URL="https://api.openveer.com"' >> ~/.zshrc
echo 'export ANTHROPIC_API_KEY="sk-xxxxxxxxxxxx"' >> ~/.zshrc
echo 'export ANTHROPIC_MODEL="claude-opus-5"' >> ~/.zshrc
source ~/.zshrc

macOS / Linux (bash)

echo 'export ANTHROPIC_BASE_URL="https://api.openveer.com"' >> ~/.bashrc
echo 'export ANTHROPIC_API_KEY="sk-xxxxxxxxxxxx"' >> ~/.bashrc
echo 'export ANTHROPIC_MODEL="claude-opus-5"' >> ~/.bashrc
source ~/.bashrc

Windows

Option A: GUI Settings

1. Right-click "This PC" → "Properties" → "Advanced system settings" → "Environment Variables"
2. Add new user variables:
   * ANTHROPIC_BASE_URL = https://api.openveer.com
   * ANTHROPIC_API_KEY = sk-xxxxxxxxxxxx
   * ANTHROPIC_MODEL = claude-opus-5
3. Restart the terminal

Option B: PowerShell Commands

[System.Environment]::SetEnvironmentVariable('ANTHROPIC_BASE_URL', 'https://api.openveer.com', 'User')
[System.Environment]::SetEnvironmentVariable('ANTHROPIC_API_KEY', 'sk-xxxxxxxxxxxx', 'User')
[System.Environment]::SetEnvironmentVariable('ANTHROPIC_MODEL', 'claude-opus-5', 'User')

Method 3: Temporary Environment Variables

Suitable for quick testing — configuration is lost when the terminal is closed.

macOS / Linux

export ANTHROPIC_API_KEY="sk-xxxxxxxxxxxx"
export ANTHROPIC_MODEL="claude-opus-5"
export ANTHROPIC_BASE_URL="https://api.openveer.com"
claude

Windows (PowerShell)

$env:ANTHROPIC_API_KEY="sk-xxxxxxxxxxxx"
$env:ANTHROPIC_MODEL="claude-opus-5"
$env:ANTHROPIC_BASE_URL="https://api.openveer.com"
claude

Windows (CMD)

set ANTHROPIC_API_KEY=sk-xxxxxxxxxxxx
set ANTHROPIC_MODEL=claude-opus-5
set ANTHROPIC_BASE_URL=https://api.openveer.com
claude

Step 3: Start Using

Verify Configuration

Launch Claude Code and send a simple message to confirm:

claude "Hello"

If you receive an AI response, the configuration is successful. If you see 401, 403, or other errors, refer to the FAQ below.

Usage Modes

Claude Code offers two interaction modes:

• Interactive mode: Run claude for continuous conversation, ideal for complex tasks
• Single command: Run claude "your question" for a one-off response, ideal for quick queries

Supported Models

OpenVeer supports the full Claude model series. You can flexibly switch based on your needs:

All models above support a Thinking (extended reasoning) version — append -thinking to the model ID, e.g., claude-opus-4-6-thinking. Thinking mode is ideal for complex tasks requiring deep reasoning.

To switch models, use the /model command in interactive mode, or modify the ANTHROPIC_MODEL field in your configuration and restart.

Common Commands

Here are frequently used commands and shortcuts in Claude Code:

FAQ

Q1: Still seeing the login selection page after configuration?

If "Select login method" still appears after startup, the configuration hasn't taken effect.

Troubleshooting steps:

1. Using settings.json: Check if the file path is correct
   * Windows: C:\Users\<username>\.claude\settings.json
   * macOS / Linux: ~/.claude/settings.json
2. Using environment variables: Make sure you launched Claude Code from the same terminal window where you set the variables
3. Check JSON format: Ensure brackets, commas, and quotes are all correct

Q2: Getting 401 / 403 errors?

Also ensure ANTHROPIC_BASE_URL is set to https://api.openveer.com, not the official Anthropic address.

Q3: "Unable to connect" error?

This means Claude Code failed to connect to the API service.

1. Check your network connection
2. Verify ANTHROPIC_BASE_URL is configured correctly
3. If using a proxy, ensure it allows access to api.openveer.com

Q4: "Auth conflict" error?

If you see a message like:

Auth conflict: Both a token (claude.ai) and an API key (ANTHROPIC_API_KEY) are set.

This means you're logged into claude.ai and have an API Key set at the same time, causing a conflict.

Solution: Run /logout in Claude Code interactive mode to sign out of claude.ai, keeping only the API Key configuration.

Q5: max_tokens error when using non-Claude models?

If you see an error like:

max_tokens is too large: 32000. This model supports at most 16384 completion tokens.

This is because Claude Code sends requests with Claude model parameters by default (32000 tokens), while some non-Claude models (e.g., gpt-4o) have lower token limits.

Solution: Switch to a Claude model (e.g., claude-sonnet-4-6), which is natively compatible with Claude Code and won't cause this issue.

Q6: Slow response?

1. Switch to a faster model (e.g., claude-haiku-4-5-20251001 or claude-sonnet-4-6)
2. Shorten your prompts to reduce context length
3. Check your network conditions

Q7: How to switch models?

Two ways:

1. In interactive mode: Enter the /model command
2. Edit configuration: Change the ANTHROPIC_MODEL field in settings.json or environment variables, then restart Claude Code

Q8: Does Claude Code automatically read local files?

No. Claude Code requires you to explicitly reference files, and it will ask for confirmation before performing sensitive operations. It's recommended to use it in dedicated project directories.

Q9: How to analyze local files with Claude Code?

In interactive mode, you can reference files by:

• Entering the file path directly
• Dragging files into the terminal window
• Copy-pasting file content

Q10: How to check usage and costs?

Log in to the OpenVeer Console to view:

• API call statistics
• Token consumption details
• Cost statistics and trends

Support

If you encounter any issues:

• 📚 OpenVeer Documentation
• 💬 Discord Community
• 🐦 Twitter @AIOpenVeer
• 📧 Technical Support: zhihong@openveer.com

This documentation is built and hosted on Mintlify, a developer documentation platform.