> ## 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.

# Add Global AI to Dify as a Model Provider

> Connect Global AI to Dify through the OpenAI-API-compatible plugin and use it as the reasoning model in workflows, agents, or applications.

Dify connects to Global AI through the built-in OpenAI-API-compatible provider, giving your apps and workflows access to every supported reasoning model.

<Steps>
  <Step title="Open settings">
    In Dify, click your avatar in the top-right corner and choose **Settings**.
  </Step>

  <Step title="Install the OpenAI-API-compatible provider">
    Under **Model Provider**, find **OpenAI-API-compatible** and click **Install**.
  </Step>

  <Step title="Add a model">
    After installing, click **Add Model** inside the OpenAI-API-compatible provider and fill in:

    * **Model name**: copy a model ID from [globalai.vip/pricing](https://globalai.vip/pricing), for example `gpt-4o`
    * **API endpoint URL**: `https://globalai.vip/v1`
    * **API Key**: paste your `sk-` token

    Leave the remaining fields at their defaults and save.
  </Step>

  <Step title="Set the system reasoning model">
    Back in **System Model Settings → System Reasoning Model**, select the model you just added and save.
  </Step>

  <Step title="Use it in an app">
    Return to your Dify app and pick the model in a workflow node or orchestration. A successful response confirms the integration.
  </Step>
</Steps>

## Add multiple models

If an app needs GPT, Claude, and Gemini side by side, repeat the **Add Model** step inside the same OpenAI-API-compatible provider. Create one record per model ID and reuse the same API key across all of them.

<Tip>
  Claude and Gemini come from different vendors, but Global AI exposes both in OpenAI-compatible format, so both work through the OpenAI-API-compatible provider. Just keep the model ID identical to the one on the [pricing page](https://globalai.vip/pricing).
</Tip>

## Troubleshooting

<AccordionGroup>
  <Accordion title="Authentication failed when adding a model">
    Confirm the API endpoint URL is `https://globalai.vip/v1` (including `/v1`) and that there are no extra spaces around your key.
  </Accordion>

  <Accordion title="Balance drops quickly during use">
    A Dify workflow can call the model several times per run, and long prompts plus retrieved context all count as tokens. Filter [Logs](/en/usage/logs) by token name to find the node consuming the most.
  </Accordion>
</AccordionGroup>


## Related topics

- [Global AI Client Integrations: Parameters and Selection](/en/clients/overview.md)
- [ChatBox with Global AI: Desktop and Mobile Setup](/en/clients/chatbox.md)
- [WorkBuddy with Global AI: Custom Provider Configuration](/en/clients/workbuddy.md)
- [Cherry Studio with Global AI: Claude, GPT, Gemini Setup](/en/clients/cherry-studio.md)
- [Contact Us: Global AI Support and Partnerships](/en/contact.md)