Added a retry button to the bottom-right corner of the chat bubble.
Fixed an issue where the backend could not properly handle interruption when the frontend stopped a streaming response.
Added a system role prompt called “S2O” to convert SillyTavern character cards into a format compatible with OSRChat.
"Bro, does it support multi-axis mode for the SR6?
Multi-axis isn’t supported for now. The main difficulty is that multi-axis behavior is hard to describe, and AI would likely struggle to understand, analyze, and generate multi-axis funscripts.
Fixed an issue where chat bubbles were being squashed.
While the model is generating text, it is displayed as small gray scrolling text until the output is complete. @dollshotz
Ill give this a try after work today, thanks!
Im not getting the Comfyui connected still fun. I have comfyui and the SXDL model etc, but its not able to make pics
edit: ok after talking and playing with it for a few it is generating prompts but not connected, so as soon as I figure out how to get it connected… anyway, its a lot of fun anyway, lol, thanks
The way text-to-image generation works is as follows: the frontend sends the prompt to the OSRChat backend, and the backend submits the corresponding ComfyUI workflow (either SDXL or ZIT text-to-image workflow) to the connected ComfyUI server based on the configuration. The ComfyUI server returns an ID to track the generation progress. If the image is generated successfully, it is downloaded from the ComfyUI server to the OSRChat backend, and then sent back to the frontend.
What needs to be ensured is that the correct models are available on the ComfyUI server with the correct names and ports configured properly. It is also recommended to use the latest version of ComfyUI and deploy it from source rather than using ComfyUI Portable. Additionally, before submitting the workflow, it’s best to preload the required models into VRAM or system memory to avoid request timeouts caused by slow model loading.
[Translated by Google Gemini]
I modified OSRChat 1.2 through vibe-coding to add multi-axis support, allowing it to fully utilize the SR6. Here are the details:
server.py to start it, or open a terminal in the OSRChat root directory and enter python server.py to run it.Finally, a huge thanks to the original author of OSRChat, @Karasukaigan, for open-sourcing it!
[Original Version]
我通过vibe-coding修改了OSRChat 1.2，让其可以支持多轴，能完全利用上SR6。以下是详细信息：
最后，十分感谢OSRChat原作者@Karasukaigan的开源！
query on how to use this, interested in having a multi axis ai experience!
This sounds very promising 
Sadly, I could not get an LLM connection up and running.
I’m trying to use Ollama with a local installation.
Neither leaving the API key blank or entering an API key does work.
It also shows “Disconnected” in the upper right area.
In browsers I can see the following error (tried Crome and Firefox).
Maybe I’m doing something wrong. It sounds like a browser security feature is blocking the OSRChat from working. But that seems unlikely 
Perhaps someone who has managed to get it working can help?
Thanks in advance! 
OSRChat is no longer maintained, so I’d recommend switching to TavLite instead.
LLM connection issues are usually caused by configuration problems, especially the Base URL or Model settings.
For Ollama, the Base URL is typically either http://localhost:11434/v1 or http://192.168.1.xxx:11434/v1 if you’re connecting over your local network.
Thanks for the headsup, I’ll have a look at Tavlite :)
