How to set up gate.ai cursor in Cursor
Gate.AI’s OpenAI-compatible integration lets developers connect Cursor to Gate.AI (as of June 2026) by using a Gate.AI API key, a Gate.AI OpenAI base URL, and supported model IDs. For developers who want Cursor Chat, Composer, Agent, or similar code interfaces to send model requests through Gate.AI instead of the default model provider, this setup is especially important. This guide only covers Cursor configuration steps. For broader API integration patterns, see Gate.AI API Integration Guide.
Prerequisites
- Cursor is installed.
- You have a Gate.AI API key.
Source: Gate.AI official documentation and the accompanying Cursor configuration screenshots, as of June 2026.
What you can do after completing this guide
After you configure gate.ai in Cursor, you can add Gate.AI as a custom model router in Cursor. You’ll enter your Gate.AI API key and an OpenAI-compatible base URL, select a model such as deepseek/deepseek-v3.2, and then use that model in Cursor’s chat or Agent interface.
This guide only covers configuration in the Cursor UI. It does not cover SDK integration, Claude Code configuration, Claude Desktop configuration, or enterprise permission settings.
Step 1: Open Cursor Settings
This step opens Cursor’s configuration area for managing models and API-related settings.
Action: In Cursor, go to the top-right corner and click the Settings icon. The Cursor entry path described in the Gate.AI official documentation is Top-right menu → Settings.
Before changing model settings, enter Cursor Settings from the top-right menu first.
Step 2: Open Model Settings
This step takes you to Cursor’s model configuration page.
Action: In Cursor Settings, select Models from the left sidebar. If you don’t see the full model list, choose View all models. The Gate.AI official documentation lists the Cursor path for adding a custom model as Models → View all models.
Use the Models section to manage built-in models, custom models, and API settings.
Step 3: Click Add Custom Model
This step creates a new custom model entry for Gate.AI.
Action: Scroll to the bottom of the model list and click Add custom model. The Gate.AI official documentation states that you must scroll to the bottom and then select Add custom model.
After adding a custom model, a field to enter the Gate.AI model ID will appear.
Step 4: Add the Gate.AI Model ID
This step tells Cursor which Gate.AI-supported model should handle your requests.
Action: In Cursor, enter the Gate.AI model ID, then click Add. Use the provider/model-name format, such as deepseek/deepseek-v3.2. The Gate.AI official documentation notes that Cursor should fetch model IDs from the model marketplace, and you should not use auto during configuration.
Per the Gate.AI official documentation (as of June 2026), model IDs use the provider/model-name format. deepseek/deepseek-v3.2 is an example of a supported model ID.
Step 5: Enter the Gate.AI API Key and Base URL
This step connects Cursor’s OpenAI-compatible API fields to Gate.AI.
Action: In the API key section, enable OpenAI API key and paste your Gate.AI API key. Then enable Override OpenAI base URL and enter:
xx/openai/v1
The Gate.AI official documentation lists xx/openai/v1 as the OpenAI-compatible base URL. It also emphasizes that the correct API path is /openai/v1, not just /v1.
Enter the Gate.AI model ID, Gate.AI API key, and the xx/openai/v1 base URL in Cursor.
Step 6: Select the Gate.AI Model in Cursor Chat
This step confirms that Cursor can use the Gate.AI model in code conversations.
Action: Close Cursor Settings. Open Cursor Chat, Composer, Agent, or another conversation interface. Then select the model you added in the model picker. The Gate.AI official documentation notes that the added model can be selected in Chat, Composer, Agent, and other conversation interfaces.
You should see the configured model name in the model picker, such as deepseek/deepseek-v3.2.
After configuration, select the Gate.AI model from Cursor’s model picker and send your instruction.
What values should be entered for the gate.ai Cursor configuration?
| Cursor field | Value to enter | Configuration rules |
|---|---|---|
| Custom model ID | Example: deepseek/deepseek-v3.2 | Use provider/model-name format; do not use auto. |
| OpenAI API key | Your Gate.AI API key | Confirm the key is valid and copied correctly. |
| Override OpenAI base URL | xx/openai/v1 | Use /openai/v1, not just /v1. |
| Model picker | The model you added | Select the model before testing your instruction. |
Model ID and base URL are the most critical fields in the gate.ai Cursor configuration. If Cursor accepts the key but can’t route requests, first check the model ID format and the base URL before changing other settings.
Gate.AI not working in Cursor? Troubleshooting checklist
Issue: Cursor shows
401or the connection fails.- Cause: The Gate.AI API key is invalid, expired, revoked, disabled, copied incorrectly, or the account balance is insufficient.
- Fix: Check the status of your Gate.AI API key, copy it again, and confirm your account balance is sufficient. The Gate.AI official documentation notes that an invalid API key and insufficient balance are common causes of authentication or request failures.
Issue: Cursor reports the model is unavailable or can’t be found.
- Cause: The model ID is misspelled, it doesn’t use the
provider/model-nameformat, or it is set toauto. - Fix: Copy the exact model ID from the Gate.AI model list and use a specific model such as
deepseek/deepseek-v3.2.
- Cause: The model ID is misspelled, it doesn’t use the
Issue: Cursor returns an unknown API path or a 404-type error.
- Cause: The base URL was entered as
https://api.gate.ai/v1or another incorrect path. - Fix: Change the value to
https://api.gate.ai/openai/v1.
- Cause: The base URL was entered as
Issue: The custom model doesn’t appear in Cursor Chat.
- Cause: You didn’t save the Cursor settings page, or Cursor hasn’t refreshed the model list.
- Fix: Save the settings, close Settings, restart Cursor, and then check the model picker again.
Issue: One model works, but another does not.
- Cause: The selected model doesn’t exactly match the Gate.AI model ID, or the model is restricted by your account or risk-control policies.
- Fix: Confirm the model ID format in the Gate.AI official documentation, and check whether the account or risk-control policies restrict that model.
What to configure or develop next?
For a broader developer integration workflow, refer to Gate.AI API Integration Guide. This guide explains the role of Gate.AI API access in OpenAI-compatible developer flows.
Use Gate.AI documentation for official configuration references. For model selection, see Gate.AI model list. If you need budget or usage review, check Gate.AI pricing before configuring in production.
FAQ
Can I use auto as the model ID in Cursor?
No. The Gate.AI official documentation states that Cursor configuration should use the specific model ID from the model marketplace and that you should not use auto.
After configuration, where do I select the Gate.AI model?
Open Cursor Chat, Composer, Agent, or another conversation interface, then select the added model in Cursor’s model picker.
Why is the API key usable elsewhere but not in Cursor?
Cursor may be using an unsaved model entry, an incorrect base URL, or an incorrect model ID. Re-check the model ID, the API key toggle, the base URL toggle, and the model selected in the model picker.
