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

# Global AI Client Integrations: Parameters and Selection

> Common parameters, model group notes, and selection guidance for integrating clients with Global AI across all client types.

Global AI is compatible with the OpenAI, Claude, and Gemini protocols, so almost every mainstream AI client can connect. This page summarizes the shared parameters and selection guidance you need before you start. See the client-specific pages below for step-by-step configuration.

<h2 id="before-you-integrate">
  Before you integrate
</h2>

Complete these two steps before configuring any client:

<Steps>
  <Step title="Add quota">
    If you already have a redemption code, go to the **Redemption code** section on [globalai.vip/wallet](https://globalai.vip/wallet), paste it, and click **Redeem**. You can also [top up online](https://globalai.vip/wallet) directly on the site.
  </Step>

  <Step title="Get an API key">
    Visit [globalai.vip/keys](https://globalai.vip/keys) and copy the default system token. If there isn't one, click **Add token** to create one. Keys start with `sk-`. See [Token management](/en/usage/tokens) for details.
  </Step>
</Steps>

## Common configuration parameters

Most clients need only these four pieces of information:

| Parameter           | Value                                               | Notes                                                                                         |
| ------------------- | --------------------------------------------------- | --------------------------------------------------------------------------------------------- |
| API host / Base URL | `https://globalai.vip` or `https://globalai.vip/v1` | Follow the client's convention. Native Claude format usually needs the host URL without `/v1` |
| API key             | Your `sk-xxxxxxxx` token                            | Copy from [/keys](https://globalai.vip/keys)                                                  |
| Model name          | See [Pricing](https://globalai.vip/pricing)         | Copy the model ID and paste it into the client                                                |
| Provider            | OpenAI, Anthropic, or Custom                        | Choose based on the model's protocol                                                          |

<Note>
  When you paste the API URL and key, check for stray whitespace at the start or end. Extra spaces cause connection failures.
</Note>

<h2 id="a-note-on-claude-model-cost">
  A note on Claude model cost
</h2>

Agent-style tools like Claude Code, Cline, and Roo Code resend the full conversation history and tool call context every round. This is especially heavy for long sessions or large codebases, and actual charges can be much higher than intuition suggests. Before you start:

* Create a dedicated token for Agent-style tools in [Token management](/en/usage/tokens) so you can track spend separately by token name in the [Usage logs](/en/usage/logs).
* Keep [Usage logs](/en/usage/logs) open while working. Filter by time and token to inspect each request's token count and quota deduction.
* If a single task consumes more than expected, trim the system prompt, shorten `CLAUDE.md` and `AGENTS.md` context files, or use a cheaper model such as Claude Haiku for lightweight tasks.

## Pick a client page by type

<CardGroup cols={2}>
  <Card title="ChatBox" icon="comments" href="/en/clients/chatbox">
    Cross-platform graphical chat client. Desktop and mobile use the same setup.
  </Card>

  <Card title="Cherry Studio" icon="cherries" href="/en/clients/cherry-studio">
    Multi-provider chat client with one-click retrieval of all available models.
  </Card>

  <Card title="ChatWise" icon="messages" href="/en/clients/chatwise">
    Runs OpenAI, Anthropic, and Gemini providers side by side in one UI.
  </Card>

  <Card title="Claude Code" icon="terminal" href="/en/clients/claude-code">
    Official Claude Code CLI setup on Windows, macOS, and Linux.
  </Card>

  <Card title="OpenAI Codex" icon="code-branch" href="/en/clients/codex">
    Integrate Codex CLI through CC Switch or by editing config files manually.
  </Card>

  <Card title="Coding Agent extensions" icon="robot" href="/en/clients/coding-agents">
    VS Code and Cursor extensions such as Cline, Roo Code, and WorkBuddy.
  </Card>

  <Card title="OpenCode / CC-Switch" icon="toolbox" href="/en/clients/opencode">
    OpenCode configuration files and one-click provider switching with CC-Switch.
  </Card>

  <Card title="Dify workflows" icon="diagram-project" href="/en/clients/dify">
    Integrate Dify through the OpenAI-API-compatible provider.
  </Card>
</CardGroup>

## Running into issues?

<CardGroup cols={2}>
  <Card title="Common error codes" icon="triangle-exclamation" href="/en/faq#tokens-and-api-calls">
    Troubleshooting for 401, 403, 429, and more.
  </Card>

  <Card title="Contact support" icon="headset" href="/en/faq">
    Working hours 09:00 - 23:00. Partnerships, invoicing, and refunds welcome.
  </Card>
</CardGroup>


## Related topics

- [OpenCode & CC-Switch with Global AI: Integration Guide](/en/clients/opencode.md)
- [Welcome to Global AI](/en/index.md)
- [Global AI FAQ: Accounts, Tokens, Calls, and Billing](/en/faq.md)
- [Global AI Token Management: Create, Limit, and Secure](/en/usage/tokens.md)
- [Cherry Studio with Global AI: Claude, GPT, Gemini Setup](/en/clients/cherry-studio.md)