>
    Strata Copilot
    Add an LLM Integration
    Focus
    Download PDF
    Focus
    1. Home
    2. Prisma AIRS
    3. AI Gateway
    4. Add an LLM Integration
    Download PDF
    Prisma AIRS

    Add an LLM Integration

    Table of Contents

    Add an LLM Integration

    Add a large language model provider integration in AI Gateway so workspaces can route requests to that provider's models.
    Where Can I Use This?What Do I Need?
    • Prisma AIRS AI Gateway (Americas region)
    • AI Gateway activated
    • API key from the LLM provider
    • Strata Cloud Manager access
    An LLM integration connects AI Gateway to a specific provider (such as OpenAI, Anthropic, AWS Bedrock, or Google Gemini) using the provider's API key. After you create an integration, you can provision it to specific workspaces, set budget and rate limits, and control which models are available. Provider API keys are encrypted at rest and are not retrievable after they are saved. For private or self-hosted models, use the custom endpoint option.
    1. In Strata Cloud Manager, select AI SecurityAI GatewayLLM Integrations and click Add Integration.
    2. Select your LLM provider from the list, or select Custom for a private or self-hosted model.
    3. In Integration Name, enter a unique name for this integration.
      Use a descriptive name that identifies the provider and intended use, for example, openai-prod or bedrock-us-east. The slug generated from this name is used in API requests to reference the integration.
    4. In API Key, enter the API key from your LLM provider.
      The key is encrypted at rest immediately after you save. You cannot retrieve the value after saving — if you need to rotate the key, edit the integration and enter the new key.
    5. (Optional) For providers that support additional authentication or regional routing, complete any additional fields displayed on the form, such as Resource Name or API Version for Azure OpenAI, or Region for AWS Bedrock.
    6. (Optional) For a custom endpoint, enter the base URL in Custom Host and add any required headers in the Custom Headers section.
    7. (Optional) To adjust how AI Gateway calculates flex credit consumption for this integration, set a Pricing Multiplier.
      The multiplier applies a scaling factor to the token-to-credit conversion. A value of 1.0 uses the standard rate (1 flex credit = 4 characters). Enter a value less than 1 to apply a discount, or greater than 1 to apply a surcharge.
    8. Click Next to proceed to workspace provisioning.
      Workspace provisioning controls which workspaces can use this integration and sets per-workspace budget and rate limits. Continue with the steps in Configure Workspace Provisioning, Configure Rate Limits, Configure Budget Limits, and Configure Model Provisioning before clicking Create Integration.
    9. Click Create Integration.
    10. Click Commit to apply the configuration.

    © 2026 Palo Alto Networks, Inc. All rights reserved.