Implements Artifacts with Local Ollama models

[CIDE-Jim]

The artifacts interface does not appear when using a local Ollama model. Is it possible to extend artifacts to be available to local models as well?

[Exioness]

Artifacts only work with LLM models that are capable of processing them, which requires the ability to follow complex instructions. Therefore, smaller models (e.g., 2B, 3B parameters) typically won't be able to handle artifacts effectively, as they lack the capacity to follow instructions in this way.

[CIDE-Jim]

Thanks for the clarification. I appreciate your insight.

[sanjeevkumarraob]

I am using llama 3.1 8b parameter model with my Ollama. I am getting the code suggestions in a regular code block but not in artifacts panel. Is there any configuration to be provided in the Ollama custom endpoint?

[danny-avila]

If you haven't seen the video, it explains how this feature works and how it can be used: https://www.youtube.com/watch?v=GfTj7O4gmd0

[sanjeevkumarraob]

Yeah. I have seen the video and pretty impressed about the functionality. at one point, the prompt prefix includes a custom definition for artifacts. But I can't find these information anywhere in the documentation. Is it possible to add more details about generating artifacts in the docs?