Delegated vision for text only models in Discourse AI

News and Events Announcements
1

We just released support for delegated vision in Discourse AI :confetti_ball:

This enables you to use text only models such as DeepSeek v4 Flash with the full feature set of Discourse AI by delegating vision to a vision capable model such as Qwen or Gemma.

example of ai bot recognizing image
example of ai bot recognizing image1299×1579 159 KB

A new option now exists under your LLM configuration:

LLM settings, where use another model for image handling is selected
LLM settings, where use another model for image handling is selected620×289 10.7 KB

Learn more about LLM configs at:

4 Likes
2

nice work Sam and ai-team, this is a cool and useful feature.

2 Likes