Skip to content

Manual and automatic client configuration

Choose the model family first, then use the exact protocol and URL required by the client.

01

Choose your key

API Credit or Token Budget

02

Choose your client

Claude Code, Codex, or another app

03

Verify the route

Run the smallest health check

Not sure a client belongs here? Supported clients lists every client CloudService covers, which protocol each one speaks, and whether the CLI can configure it for you.

Automatic configuration

Install the CloudService CLI once and use its direct command for supported clients:

bash
npm install -g cloudservice@latest
cloudservice

The CLI identifies the key's provider and billing mode, offers only compatible clients and models, backs up existing files, and verifies the resulting route. OpenCode, Cursor, Roo Code, and Cline for VS Code use the normalized OpenAI-compatible chat route, so they can be configured with a Claude-only key as well as an OpenAI key. Cline CLI uses the Responses route. No npx command is required. Run cloudservice doctor at any time to check the current setup.

Choose the URL before manual setup

Billing mode and transport are separate decisions: API credit versus an API-token key determines the billing surface; OpenAI-compatible versus native Anthropic Messages determines whether the client field ends in /v1.

Start with the key you bought

Choose one billing path

Your choice is remembered across the documentation. Then select the card that matches your client protocol.

For a key funded with a dollar credit balance.

OpenAI-compatible Chat Completions or Responses

Use with: Codex, OpenCode, Cursor, Roo Code, SDKs, and direct HTTP

Base URL to enter

https://api.yourdomain.example/v1

Native Anthropic Messages

Use with: Claude Code, Claude Desktop, and native Messages clients

Base URL to enter

https://api.yourdomain.example

The protocol decides whether the client field ends in /v1. OpenAI-compatible clients use a base URL ending in /v1. Native Anthropic clients use the origin only, without /v1, because they append /v1/messages themselves. A Claude token key therefore reaches /token/v1/messages when the native client builds its request.

Use the key's matching billing surface and only models returned by its authenticated catalog. A Claude token key and an OpenAI token key are provider-scoped; neither can be swapped onto the other's models.

Claude Desktop is a native Anthropic client with its own third-party inference settings. Follow the Claude Desktop guide; it does not read Claude Code's configuration.

Choose Claude Code or Codex

These are separate manual configurations. Choose the client you are actually opening; do not copy one client's settings into the other.

Claude Code

Native Anthropic Messages. Enter an origin URL with no trailing /v1.

Open Claude Code setup →

Codex

OpenAI Responses. Enter a base URL ending in /v1 and set wire_api = "responses".

Open Codex setup →

Clients that can expose either model family

These clients need a separate model-family choice inside the guide. Their compatibility adapters are not the same thing as the native Claude Code or Codex client:

  • OpenCode — OpenAI-compatible custom provider.
  • OpenClaw — separate Anthropic Messages and OpenAI Responses providers.
  • Hermes — named OpenAI-compatible chat-completions provider.
  • Roo Code for VS Code — OpenAI Compatible profiles for Claude and Codex; never the OpenAI Codex OAuth provider. It is not offered for DeepSeek V4 Flash.
  • Cline — Cline CLI uses Responses; Cline for VS Code uses OpenAI-compatible chat with one raw key and no custom authentication header.
  • Continue for VS Code — automatic or manual Chat/Edit profile for deepseek-v4-flash.
  • Cursor — experimental custom OpenAI-compatible model support.
  • n8n — AI Agent workflows through the OpenAI Chat Model sub-node, for Claude, Codex, and DeepSeek; includes an importable starter workflow.

Other client guides

  • Claude Desktop on 3P — managed third-party inference settings, separate from consumer Claude.ai chats and Claude Code. Enter the native Anthropic origin with no /v1.
  • TRAE SOLO — custom OpenAI-compatible model settings.
  • Cherry Studio — OpenAI-compatible desktop provider.
  • API Code — OpenAI SDKs or direct HTTP.

Billing URL rule

API-credit and token-pack keys are deliberately isolated. Use the URL shown for the purchased key and client protocol, and expose only the models returned by its authenticated catalog.