Choose a protocol first
Check the protocol selected in your tool’s guide before entering a Base URL. Anthropic Messages useshttps://bettertoken.ai; OpenAI-compatible examples use https://bettertoken.ai/v1. The GitHub Copilot Chat Completions example and the TRAE full-URL example use the complete https://bettertoken.ai/v1/chat/completions address. In TRAE, check the URL mode available in your region and client version.
The protocol comparison explains the difference. The protocol follows the client: Claude Code can also call GPT models through BetterToken’s Anthropic Messages endpoint.
Prerequisites
Create a BetterToken API Key using the quickstart, confirm your balance in the Dashboard, and copy an exact Model ID from the model plaza. One BetterToken API Key works with all available models; you can create separate keys for individual projects.Supported tools and setup methods
One-command setup means a BetterToken configuration script, not a command that only installs the tool. For manual setup, open the tool’s guide for its UI or configuration file. The table shows the protocols and URL fields used by these guides’ examples, rather than a complete list of each client’s protocol options.Start with your workflow
Terminal agents
Start with Claude Code, or use Codex CLI, OpenCode, Pi Agent, or Aider from the table.
IDE extensions
Configure Cline in VS Code, or open the guide for your existing editor or extension.
Platform automation
Connect an n8n workflow or use Dify for an AI application.
Configuration management
Manage connection settings for several clients with CC Switch.

