There should be options other then chatgpt for inference.

[coderwithsense]

Plan:

• Add anthropic, google gemini, grok, ollama, kimi k2.5 etc

[github-actions]

Found 2 possible duplicate issue(s):

1. Support custom models #90
2. Ollama #27

This issue will be automatically closed as a duplicate in 3 days.

• To prevent auto-closure, add a comment or react with 👎 on this comment.

[KylinMountain]

@coderwithsense we will support these via litellm or anything else to custom llm provider

[coderwithsense]

I can help!

[coderwithsense]

@KylinMountain @rejojer I am thinking --model will take model name, it will check if that model is available to use (pydantic), a few new functions in utils, and the usage accordingly everywhere.
opt will not change, it will remain same.
Can i start implementing this, or you can add something which i need to keep in mind before doing this.
Using litellm.

[KylinMountain]

This is now supported — multi-provider inference landed via the LiteLLM integration (#168). Anthropic (anthropic/claude-...), Gemini (gemini/...), Ollama (ollama/...), Grok (xai/grok-...), and many others work just by setting the model name; full list: https://docs.litellm.ai/docs/providers

Thanks @coderwithsense for offering to help and for pushing on this early! Closing as implemented.