← all conversations

AI Generative Tool

2025-12-252 turns2,143 charsgpt-5-2
social-mediadevelopercode-sharing

Summary

User wants to share a GitHub project link with developers and vibe coders after being banned from social media.

Messages

make shorter: Hey, I was banned from ALL of social media because I am mean so I have nowhere to share this, but I worked hard making this and maybe some of y'all are developers or at least vibe coders and can appreciate: https://github.com/kliewerdaniel/ConCreat It is a Next.js frontend using Hunyuan for video generation, Z-Image-Turbo for image generation, Ollama for text generation and Chatterbox for audio generation. I tested it on a Macbook Pro M4-Pro 48GB setup. It all runs locally so it doesn't need the internet, runs entirely for free, and no data leaves the computer. I built it to centralize all of the local generative AI projects I have cloning voices for text to speech or using any LLM I want with ollama or generating any photo or video. You have to run ComfyUI in the background, but I integrated the workflows and have links to all the files you have to download. It takes up a lot of space, but once it is installed you can generate all you want. I hard coded in the workflow 200x200 5seconds for the video, so I am going to add that to input to the UI next, but that was the setting which generates in around 4 minutes, the images generate in around 90 seconds and the audio takes a bit as well. So it is not the fastest setup in the world. But if you like tinkering with AI like I do, maybe you will like this.
Here’s a much shorter, cleaner version while keeping the vibe: > I got banned from basically all social media, so I have nowhere to share this 😅 but I built something I think devs / vibe coders might appreciate: > https://github.com/kliewerdaniel/ConCreat > > It’s a local-only Next.js frontend that unifies text (Ollama), image (Z-Image-Turbo), video (Hunyuan), and audio (Chatterbox) generation. Everything runs fully offline, free, and no data ever leaves your machine. > > I tested it on a MacBook Pro M4-Pro (48GB). You do need ComfyUI running in the background, but all workflows and downloads are wired up. It’s not blazing fast (video ~4 min, images ~90s), but once installed you can generate as much as you want. > > If you like tinkering with local AI systems, you might enjoy this.
gpt-5-2