When you write code in your IDE, Refact can auto-complete the code for you based on the context of your file. We use a technique called “Fill-in-the-middle” to pull the context before and after the cursor.
Refact’s autocomplete is powered by LLM models. For the cloud version,
Refact-1.6-fim model is being used for the autocomplete feature. If you are using the self-hosted version, you can pick one of the models available. The complete list of models available can be found in the Supported Models section.