Supporting Adapters in CodeCompanion

[olimorris]

I wanted to share my thoughts regarding supporting additional adapters in CodeCompanion.

As the plugin continues to grow in popularity, so do the requests for adding support for additional LLMs and adapters. This is a good thing and to be expected. After-all, with every week we have new entrants to the LLM world which grab the headlines.

However, it's impossible to support every LLM in the plugin. Doing so will add additional maintenance with potentially not a whole lot of gain for the wider user base. Anticipating this 10+ months ago, I introduced a flexible adapter system and created the guide on making your own Adapters.

My stance

As of today, I'll be happy to accept PRs for any LLMs that sit on the first page of the ProLLM Leaderboard.

However, this assumes:

1. The LLM cannot be used via the ollama adapter and is hosted by a third-party
2. The LLM is not OpenAI Compatible and hence can't use the plugin's openai adapter
3. Or; It's not possible for this LLM to work with the current adapter system

There will no doubt be exceptions to this rule such as with Copilot. In those cases, let's open up a discussion.

Finally, if you know of a better resource than the ProLLM dashboard, please let me know.

[olimorris]

I'm getting a number of request to add support for additional adapters in the plugin and I should have been more explicit in the post above.

If an adapter can leverage the OpenAI adapter, then I will not be accepting a PR for it. Instead, I would ask the community to share it in the discussions and then we can link to it in the documentation.

[olimorris]

I've decided to stop supporting the openai_compatible adapter as of 2025-07-29. I'll continue to keep it in the plugin as-is and also accept sensible PRs but my intention is for it to be a template that people can copy from when they're trying to connect to OpenAI compatible LLMs