KO | EN

GitLite

Minimalist Open Source Repository Search Engine

gitlite — search

#dsh #dsh-plugin #typescript #developer-tools #claude-code #ai-agent #codex #deepseek-harness #ai-agents #cli #react #mcp

comfyconnection

★ 4 Open GitHub ↗

A Unity Editor integration example for sending prompts to ComfyUI and receiving generated images via WebSocket. Useful as a starting point for building generative AI tools with Unity.

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# 🧩 Comfy Connection - Unity–ComfyUI Integration Example

![Comfy Connection splash image](https://raw.githubusercontent.com/o-l-l-i/comfyconnection/main/images/comfyconnection_splash.png)

**Comfy Connection** is a Unity Editor extension that demonstrates how to connect to a local [ComfyUI](https://github.com/comfyanonymous/ComfyUI) server to generate images using generative AI (e.g. Stable Diffusion models).

This example project shows how you can:
- Send prompts and generation parameters from Unity
- Communicate with ComfyUI via HTTP and WebSocket
- Receive generated images and handle them in the Editor

It’s intended as a starting point for building Unity tools that leverage generative AI workflows powered by ComfyUI.

---

## ✅ Features

- Unity Editor window for inputting prompt parameters
- Sends prompt to ComfyUI over HTTP
- Listens to WebSocket for progress and image results
- Displays the generated image inside the Editor
- You can optionally save images as PNG

![Comfy Connection in action](https://raw.githubusercontent.com/o-l-l-i/comfyconnection/main/images/comfyconnection_in_action.gif)

---

## 🧰 Requirements

- **Unity 6000.0.20f1** or later
- A local installation of [ComfyUI](https://github.com/comfyanonymous/ComfyUI)
- A valid SDXL-compatible `.safetensors` model (configured in script)
- [`NativeWebSocket`](https://github.com/endel/NativeWebSocket) by [endel](https://github.com/endel)

---

## ⚠️ Compatibility Note

This example is built and tested against a **vanilla install of ComfyUI**.
Custom extensions such as **ComfyUI-CrysTools** may cause issues due to differing JSON structures or behaviors.

---

## 🚀 Getting Started

1. Add the example code to an empty Unity project.
2. Add [`NativeWebSocket`](https://github.com/endel/NativeWebSocket) to your project via UPM or by copying source files.
3. Make sure your ComfyUI server is running locally.
4. In Unity, navigate to: **Window → Comfy Connection**

![Comfy Connection menu location](https://raw.githubusercontent.com/o-l-l-i/comfyconnection/main/images/comfyconnection_menu_access.png)

5. Enter prompts and parameters, click Generate Image, and observe results.

## 📦 Workflow
This project builds and sends a prompt JSON like this:
- Load checkpoint
- Encode text with CLIP
- Generate latent image
- Decode with VAE
- Save image via WebSocket

Results are parsed in the WebSocket listener:
- If the message is JSON → updates progress
- If the message is binary → loads image into Texture2D

## 🔍 What This Is (and Isn't)

This is:
- A working example for Unity developers
- A guide to ComfyUI integration over HTTP + WebSocket
- A minimal Editor UI for exploration

This is not:
- A full-featured AI image generation plugin
- A polished, production-ready tool

## 🙏 Credits
- [ComfyUI](https://github.com/comfyanonymous/ComfyUI)
- [`NativeWebSocket`](https://github.com/endel/NativeWebSocket) by [endel](https://github.com/endel)

## 📝 License
This example is released under the MIT License.
You’re free to modify, extend, and build upon it.

comfyconnection

Explore Similar Repositories

Was this content helpful?

Recent Feedback

Download README