ComfyUI Image Editing + Gemini - WebUI - Cloud serverless GPU/VPS

AI-powered image editing service using Qwen-based workflows on RunPod and Google Gemini for prompt enhancement.

�️ Interfaces

Interface	Description	Status
Image Editor	Main Interface. Interactive AI image editing using Qwen workflow.	✅ Production
Content Pipeline	"Conveyor Belt" prototype for batch content creation (Idea -> Video).	🚧 Prototype
Studio Dashboard	Classic chat & generate interface with Gemini integration.	✅ Stable

📸 Screenshots

✨ Operations

• AI Editing: Replace/modify objects in images using natural language.
• Prompt Engineering: Use Gemini to expand simple ideas into professional prompts.
• Cloud Backend: Serverless GPU execution via RunPod (no local GPU required).
• Production Ready: Built-in rate limiting, logging, and PM2 support.

🚀 Quick Start

1. Install Dependencies

   npm install

2. Configure Environment

   cp .env.example .env
   # Set GEMINI_API_KEY, RUNPOD_API_KEY, RUNPOD_ENDPOINT_ID

3. Run Locallly

   npm run dev
   # Open http://localhost:3000

📦 Deployment (VPS)

1. Setup PM2: npm install -g pm2
2. Start Server: pm2 start ecosystem.config.js
3. Docs: See DEPLOYMENT.md for full Nginx/HTTPS setup.

� Configuration

Required .env Variables:

• GEMINI_API_KEY: Google AI Studio key.
• RUNPOD_API_KEY: API Key from RunPod.
• RUNPOD_ENDPOINT_ID: Endpoint ID for your deployed ComfyUI workflow.

Default Port: 3000 (Set PORT to change)

🏗️ Architecture

graph TD
    subgraph Client
        UI[Web UI / Browser]
        Ed[Image Editor]
    end

    subgraph Server_Layer
        API[Express Server]
        Auth[Auth Middleware]
        Job[Job Manager]
    end

    subgraph AI_Services
        Gemini[Google Gemini API]
        RunPod[RunPod Serverless GPU]
    end

    UI --> |/api/generate| API
    UI --> |/api/chat| API
    
    API --> Auth
    Auth --> Job
    
    Job --> |Enhance Prompt| Gemini
    Job --> |Generate/Edit| RunPod
    
    RunPod --> |Images| Job
    Gemini --> |Text| Job
    
    Job --> |JSON Response| UI

Loading

🔄 Workflow: Image Editing

sequenceDiagram
    participant User
    participant WebUI
    participant Server
    participant Gemini as Gemini AI
    participant RunPod as ComfyUI (RunPod)

    User->>WebUI: Enters Prompt & Uploads Image
    WebUI->>Server: POST /api/generate (edit)
    
    par Prompt Enhancement
        Server->>Gemini: Enhance Prompt
        Gemini-->>Server: Optimized Prompt
    end
    
    Server->>RunPod: Submit Job (Image + Prompt)
    RunPod-->>Server: Job ID
    Server-->>WebUI: Job ID (Pending)
    
    loop Polling
        WebUI->>Server: GET /api/status/:id
        Server->>RunPod: Check Status
        RunPod-->>Server: Status / Result
        Server-->>WebUI: Status update
    end
    
    RunPod-->>Server: Final Image
    Server-->>WebUI: Display Result

Loading

📝 License

MIT