Private AI End Point

A private AI endpoint serves as an interface between users or applications and AI models, allowing them to send requests for processing and receive responses in real time. This section provides detailed steps on how to create an private AI endpoint, including selecting the appropriate model, configuring settings (such as input/output parameters and authentication), and managing access controls.

Tip 💡 To create a private AI endpoint, the workspace must be GPU-enabled to support the computational demands of AI models.

• Create private AI End Point

  To create a private AI endpoint, follow these three simple steps:

  • Click the Create button on the right side of the screen, then select Private AI Endpoint.

    

  • Fill in the required details, including the private AI Endpoint name, select a GPU-enabled workspace, choose the model repository, and enter your Hugging Face token (Refer). Then, click Next.

    

    1. Choose an instance size, configure the policy guardrails (check/uncheck as needed), and specify the minimum and maximum number of instances.

       

    2. Click Next.

    3. Review your details and click Submit.

       

• Logs and Configuration

  The private AI Endpoint Logs provide a detailed record of all activities and events related to the operation of a private AI endpoint. These logs are essential for monitoring the health and performance of the AI system, as they capture valuable information such as request responses, errors, system resource usage, and other significant events.

  

• Configuration

  The private AI Endpoint Configuration section displays key settings such as the instance size, number of instances, and policy guardrails. These guardrails include features like Prohibited Words, PCI Masking, and Prompt Injection, which help ensure secure and compliant operation of the private AI endpoint.

  

• AI Endpoint

  you can take a demo in just two simple steps:

  1. Enter access token : You need to create an access token for authentication purposes. (Refer to create the Access Token).

  2. Make requests : You can now make requests to your private AI Endpoint.

  

• Template for AI End Point

  We also provide ready-to-use templates for making requests to your private AI Endpoint, available in Go, Python, and cURL. These templates make it easy to quickly test and interact with your private AI Endpoint by demonstrating the correct request structure and required parameters for each language. Simply copy the template for your preferred language, insert your access token and endpoint URL, and you’ll be ready to make a request to your deployed private AI Endpoint.