Skip to content

[FEATURE] Allow injection of custom headers for custom models via Litellm #466

@Luyansi3

Description

@Luyansi3

Is your feature request related to a problem? Please describe.

When using Litellm for a custom model with a custom API BASE, I need to provide personalized headers in the HTTP requests. Some API providers require multiple API keys (e.g., one for the API Gateway and another for the underlying model), so it's necessary to be able to add extra headers easily when making requests via Litellm in Strix.

Describe the solution you'd like

Add a feature allowing the user to configure and inject custom headers when making a call to a model via Litellm, to handle use-cases where multiple keys or information are required in the headers.

Describe alternatives you've considered

Currently, I have to manually patch the code or create a workaround on the client side, which is not sustainable and does not fit all use-cases.

Additional context

This need appears in contexts where models are hosted on platforms that require several distinct authentications (API Gateway signature + model key, etc.), or for compliance/tracing requirements in some enterprise environments.

Metadata

Metadata

Assignees

No one assigned

    Labels

    enhancementNew feature or request

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions