> ## Documentation Index > Fetch the complete documentation index at: https://docs.llmeasy.ru/llms.txt > Use this file to discover all available pages before exploring further. # How can I use Codex and other third-party models in Claude Desktop? > Use a CC Switch local route to run GPT Key group models in Claude Desktop. Recent Claude Desktop versions changed how third-party API gateways are configured. In Gateway mode, Claude Desktop validates model IDs more strictly, so non-official model IDs such as GPT, DeepSeek, or GLM can be blocked. A common symptom is this warning at the top of the page: `Your provider setup needs a fix Some required fields are missing or malformed. Open Setup to finish configuring it.` At the same time, the third-party model selector in the lower-left area is empty, and Claude Desktop only shows a **Legacy Model** state. Claude Desktop showing Your provider setup needs a fix while the third-party model selector is empty. If you want to keep using Codex or other third-party models in Claude Desktop, route the requests through CC Switch: configure LLMEasy under **Claude Code** in CC Switch, then enter the CC Switch local service address in Claude Desktop Gateway. This path is for API keys from the **GPT Key group**. Do not enter `https://www.llmeasy.ru/v1` directly in Claude Desktop Gateway. ## Setup steps Download the latest Claude Desktop app from the official link: [Download Claude for Mac](https://claude.ai/api/desktop/darwin/universal/dmg/latest/redirect) After installation, do not sign in to a Claude account. Continue to the next step and enable developer mode. Open Claude Desktop, then choose **Help** → **Troubleshooting** → **Enable Developer Mode** from the macOS menu bar. Select Enable Developer Mode from the Claude Desktop Help menu. In the confirmation window, click **Enable**, then restart Claude Desktop. Confirm developer mode in Claude Desktop. After restarting, the macOS menu bar shows a **Developer** menu. Choose **Developer** → **Configure Third-Party Inference...**. Choose Configure Third-Party Inference from the Developer menu. Open CC Switch, switch to **Claude Code**, and click **Add Provider**. The Claude Code page in CC Switch, with the add button in the top-right corner used to create a new provider. * **Provider Name**: `LLMEasy` * **Base URL**: `https://www.llmeasy.ru` * **API Key**: your LLMEasy API Key The screenshot labels map to: 1. **Provider Name** 2. **API Key** 3. **Base URL** The provider edit page in CC Switch, showing where to enter Provider Name, API Key, and Base URL. Open **Advanced Options** and set **API Format** to **OpenAI Responses API**. The advanced options section in CC Switch, with API Format set to OpenAI Responses API. In **Model Mapping**, click **Fetch Model List**. Explicitly select dropdown values for **Primary Model**, **Thinking Model**, **Haiku Default Model**, **Sonnet Default Model**, and **Opus Default Model**. The model mapping section in CC Switch, showing Fetch Model List and the mappings for Primary, Thinking, Haiku, Sonnet, and Opus models. Use model IDs from the current **GPT Key group** models in the LLMEasy model plaza. After saving, return to the provider list: 1. Set the LLMEasy provider you just saved as **In use** 2. Enable the **CC Switch proxy** from the top-left corner After saving, return to the provider list, set LLMEasy as active, and enable the CC Switch proxy toggle in the top-left corner. Return to the CC Switch home screen, then click **Settings** in the top-left corner. The CC Switch home screen with the Settings button highlighted in the top-left corner. Open **Routing** settings. Confirm that the highlighted main routing switch and **Claude** route switch are both on. Then copy the **Service address** shown below. The CC Switch routing settings page, with the main route switch, Claude route switch, and service address copy button highlighted. Return to the Claude Desktop third-party inference settings window and select **Gateway**. Paste the CC Switch **Service address** into **Gateway base URL**, then enter your LLMEasy API Key. Set **Gateway auth scheme** to `bearer`. The Claude Desktop Gateway settings page with the CC Switch local service address entered. Add the models you want to show in **Model list**. Each model entry needs both **Model ID** and **Display name**. For example: | Model ID | Display name | | --------------- | --------------- | | `claude-opus-4` | `Claude Opus 4` | After filling in the list, click **Apply locally**. The Claude Desktop Model list with Model ID and Display name filled in, and Apply locally available. After Claude writes the local configuration, it asks you to save and restart. Click **Save & Restart** so the new gateway configuration takes effect. Claude Desktop asks to save and restart, with the Save & Restart button highlighted. After restarting, wait about 3 seconds for the sign-in choice to appear. Claude Desktop shows the Continue with Gateway sign-in option. Choose **Continue with Gateway**. After Claude opens, **Cowork 3P | Gateway** appears in the lower-left corner and you can start using it. Claude Desktop is ready after signing in with Gateway. ## Troubleshooting If you still see **Your provider setup needs a fix**, check these items first: 1. The LLMEasy provider in CC Switch is set as **In use** 2. **API Format** is set to **OpenAI Responses API** 3. **Model Mapping** explicitly selects GPT Key group models for Primary, Thinking, Haiku, Sonnet, and Opus 4. **CC Switch proxy**, the main routing switch, and the **Claude** route switch are all enabled 5. Claude Desktop **Gateway base URL** uses the local service address copied from CC Switch 6. **Model list** includes both **Model ID** and **Display name** for each model