Skip to content

Add DeepInfra as an official LLM provider #15709

Description

@ZinkDifferent

Feature Description

Add DeepInfra as a first-class supported provider in Hermes Agent, alongside Ollama, OpenRouter, and Google.

Motivation

DeepInfra offers highly competitive pricing and low-latency access to a wide range of open-weight models (Llama 3, Qwen, Mistral, etc.) using a stable OpenAI-compatible API. While it can currently be used via the custom_providers configuration, having it as a native provider would simplify setup (especially for new users) and allow for better integration with the credential pool and model discovery.

Proposed Solution

  • Add deepinfra to the PROVIDER_REGISTRY in hermes_cli/auth.py.
  • Define the default base_url as https://api.deepinfra.com/v1/openai.
  • Ensure DEEPINFRA_API_KEY is recognized as the standard environment variable.
  • (Optional) Add a dedicated setup flow in hermes setup provider.

Alternatives Considered

Continuing to use custom_providers, which requires manual JSON configuration in config.yaml and doesn't provide the same "out of the box" experience as supported providers.

Metadata

Metadata

Assignees

No one assigned

    Labels

    P3Low — cosmetic, nice to havearea/configConfig system, migrations, profilestype/featureNew feature or request

    Type

    No type

    Fields

    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