Add auth options in LLM config for local models

[metalshanked]

Is your feature request related to a problem? Please describe.
Currently, the below two options are available for local self hosted LLMs.
Ollama can be proxied via other gateways like LiteLLM which support API Keys and other auth methods

    'default': {
        'ollama': {
            'server_url': '',
            'model': '',
        },
    },

Describe the solution you'd like
Would be great if auth or even http header options are added to the self hosted LLM configs

Describe alternatives you've considered
Jerry rig whitelisting of sources

[jkppr]

Hi @metalshanked,
you can achieve this type of access to a specific type to LLM provider by adding a new provider file into the llms/providers folder. I would recommend to take inspiration from the existing ollama.py provider file and extending it with your required connection solution.

For easy testing you can use a dev setup as described here: https://timesketch.org/developers/getting-started/

Once you have a working provider to your specifications, we welcome contributions via PR to include it into the next release version.