Runchat Updates: November 14, 2025

🔤 ADD TEXT TO IMAGES (AND OTHER CANVAS IMPROVEMENTS)

• double click an image to open the editor, then use the text tool to add text tags
• this is really useful for spatializing prompts
• we've also added support for layers in the canvas editor so you can easily re-edit past changes and move your edits around
• Demo of the canvas: https://www.youtube.com/watch?v=coHf4gqS6N0

🐟 QWEN LORA INTEGRATION

• We now support adding custom Qwen Edit loras
• There are lots of these going around at the moment that are great, in particular for applying textures to objects, blending objects with a scene or changing camera angles
• Install loras directly from civit or huggingface urls at runchat.com/dashboard/training

🧠 UX IMPROVEMENTS TO NODES

• Parameter handles are now located next to their input UI to make it clearer which handle relates to what data
• If you're viewing output parameters or the node is really tiny then they default to the same behaviour as before and are neatly stacked on the side of the node
• It's now easier to add and remove inputs from the input node, you can select whatever you want to delete and then remove them all at once
• Sliders look simpler and cleaner
• Parameters appear on a node in alphabetical order so you have a little more control with how the input UI looks

• We've added support for Qwen Image Edit that gives you fine grain control over the camera angle of a subject
• This is a) fast and b) pretty reliable.
• Works with multiple inputs to quickly generate a bunch of different views

What are we working on now?

• Templates: save all inputs to a node as a reusable template to share publicly or use in other workflows.
• Template Gallery: browse dozens (and eventually 100s) of premade templates to help discover new ideas and tools