Skip to main content
TR

LLM Keys

LLM Keys allow you to use the LLM models available in SkyStudio through external applications and services. With a generated LLM API key, you can send API requests to supported models and use model responses within your own applications.

Note: The LLM Keys feature is available in the legacy SkyStudio interface. From the new interface, you can switch to the legacy interface using the Old UI option at the top of the page.

1. Accessing LLM Keys​

After switching to the legacy SkyStudio interface, follow these steps to access the LLM Keys section:

  1. Click your user profile in the bottom-left corner.

  2. Select Workspace Settings from the menu.

  3. Open LLM Keys from the left-hand menu on the Workspace Settings page.

    On the LLM Keys page, you can create LLM API keys and configure usage limits for each key.

2. Creating an LLM Key​

To create a new LLM key, enter a descriptive name for the key in the LLM API Keys field.

Before creating the key, you can configure the following settings to define request usage limits:

  • Timeout (seconds): Specifies the maximum amount of time to wait for an API request to complete.
  • Limit / second: Specifies the maximum number of requests that can be sent using the key within one second.
  • Limit / hour: Specifies the maximum number of requests that can be sent within one hour.
  • Limit / day: Specifies the maximum number of requests that can be sent within one day.

After entering the required values, click Create LLM Key.

The generated LLM API key can be used to authenticate API requests.

Important: The complete API key is only displayed during the session in which it is created. Copy the key after creating it and store it securely.

To modify the usage limits assigned to a key, update the relevant values and click Refresh.

3. LLM API Documentation​

The LLM API Documentation section explains how to use your generated LLM key to send API requests to the LLM models available in SkyStudio.

In this section, you can select the provider and model you want to use, configure model-specific settings, and review API request examples that are automatically generated based on your selections.

Integration Note​

The Integration Note, located directly below the LLM API Documentation section, provides important information about the Base URL and endpoint structure used for API integrations.

For Cursor integrations, use the following Base URL:

https://skystudio.skymod.tech/api/workflow/llm

Important: Enter this address directly in the Cursor Base URL field without adding /v1. Cursor may automatically append /v1/models, /v1/chat/completions, or /v1/responses depending on the request.

For direct HTTP requests, append the appropriate API endpoint to the Base URL. For example, use the following endpoint for the Responses API:

https://skystudio.skymod.tech/api/workflow/llm/v1/responses

3.1. Cursor Setup​

To use your generated LLM API key with Cursor, follow these steps:

  1. Create a new LLM key and copy the complete key during the session in which it is created.
  2. In Cursor, navigate to Settings → Models → API Keys and paste the key into the OpenAI API Key field.
  3. Enable Override OpenAI Base URL.
  4. Enter the following address in the Base URL field:
https://skystudio.skymod.tech/api/workflow/llm
  1. Select or add a model ID supported by the gateway. For example:
gpt-5.6

Note: Do not append /v1 to the Base URL in Cursor. Cursor may automatically append the required endpoint paths to the Base URL.

The gateway supports /v1/models, /v1/chat/completions, and /v1/responses requests, as well as function/custom tool calls and multi-step tool results.

Custom API keys configured in Cursor do not replace Cursor-specific models used for features such as Tab Completion.

3.2. Provider and Model Selection​

To begin configuring an API request, select the LLM provider you want to use from the Provider field.

Then, select the model to be used for the API request from the Model field.

Available model settings may vary depending on the selected provider and model. Additional settings such as Reasoning and Max. Completion may be displayed for supported models.

Changes made to these fields are automatically reflected in the API request examples shown further down the page.

3.3. User Message​

Enter the message or instruction you want to send to the model in the User Message field.

The entered message is automatically included in the generated API request. This allows you to review the request structure generated for the selected model and configuration before implementing the actual integration.

3.4. Endpoint and Request Information​

After selecting a provider and model, the connection and request details required for the API call are displayed.

This section includes:

  • The Endpoint to which the API request will be sent,
  • The HTTP method to be used,
  • Required HTTP headers,
  • The authentication method,
  • Required fields,
  • Available optional parameters.

The LLM API key is passed as a Bearer Token in the Authorization header.

Example:

Authorization: Bearer <YOUR_LLM_API_KEY>
Content-Type: application/json

Required and optional fields may vary depending on the selected provider and model. When implementing an integration, refer to the current fields displayed in the documentation interface for your selected configuration.

3.5. API Request Examples​

After entering the provider, model, and user message information, SkyStudio automatically generates API request examples based on the selected configuration.

The JSON Payload (Automatic) and cURL (Automatic) sections are available at the bottom of the page.

The JSON Payload (Automatic) section displays the JSON body that will be sent with the API request. Changes to the provider, model, user message, or model settings are automatically reflected in the JSON payload.

The cURL (Automatic) section provides the equivalent request in cURL format. This example includes the endpoint, HTTP method, required headers, and request body.

Use the Copy option available in either section to copy the generated example and adapt it for use in your own application or integration.

3.6. Example Response​

The Example Response section displays an example of the response structure that may be returned by the API request.

This section allows you to review how model-generated content is returned within the API response.

Request and response structures may vary depending on the selected provider, model, and model configuration. When developing an integration, use the current examples generated in the LLM API Documentation interface for your selected configuration.

3.7. Quick Troubleshooting​

If you encounter an error while connecting through Cursor or another client, first verify the Base URL configuration.

In Cursor, manually appending /v1 to the Base URL may result in an invalid URL if Cursor appends the required endpoint path again.

If you encounter an error such as Resource not found:

  1. Check the Override OpenAI Base URL setting in Cursor.
  2. Make sure the Base URL is configured exactly as follows:
https://skystudio.skymod.tech/api/workflow/llm
  1. Verify that the Base URL does not end with /v1, /v1/responses, or /v1/chat/completions.
  2. Confirm that the model ID you are using is supported by the gateway.

Note: If the issue persists, refer to the Quick troubleshooting section in the LLM API Documentation interface for the latest configuration guidance.

4. Using an LLM Key​

You can use your generated LLM key to authenticate requests sent to the SkyStudio LLM API from external applications or services.

When creating an API request, include the key as a Bearer Token in the Authorization header:

Authorization: Bearer <YOUR_LLM_API_KEY>

You can use the JSON Payload or cURL examples provided in the LLM API Documentation section to build your request. After selecting the provider and model you want to use, copy the generated example and adapt it to your application.

For direct requests to the Responses API, send the request to the following endpoint:

https://skystudio.skymod.tech/api/workflow/llm/v1/responses

5. Managing LLM Keys​

Generated LLM keys can be managed from the LLM Keys page.

Each key displays its current Status and Last Used information. Keys that are available for use are displayed as Active. If a key has not yet been used, the Last Used field may display -.

Use the eye icon next to a key to view it and the copy icon to copy it.

To change the request limits assigned to a key, update the relevant values and click Refresh to apply the changes.

If an LLM key is no longer required, click Revoke next to the corresponding key.

Revoking a key that is currently in use will cause API integrations relying on that key to stop working. Before revoking a key, make sure it is no longer being used by any applications or integrations.

Important: Store your LLM API key securely and do not share it with unauthorized users. Never expose the key in public source code or use it directly in client-side applications accessible to end users.