GenSelfie

📖 View the Full Setup Guide →

A web application that allows influencers to let their fans generate AI selfies with them. Fans can upload a photo or fetch their profile picture from social media, pay via promo code or payment, and receive an AI-generated selfie with the influencer.

Features

• Fan-facing page: Fans can enter their social media handle or upload a photo, pay via promo code/Stripe/Lightning, and generate a selfie
• Admin panel: Influencers can configure branding, pricing, upload reference images, and manage promo codes
• Social media integration: Fetch profile pictures from Twitter/X, Bluesky, GitHub, Mastodon, and Nostr
• Payment options: Promo codes, Stripe, and Bitcoin Lightning (via LNbits)
• ComfyUI backend: Image generation is handled by a ComfyUI server
• Presets: Configure multiple generation presets with different influencer images, dimensions, prompts, and pricing

Prerequisites

ComfyUI Server (RunPod)

This application requires a ComfyUI server for image generation. You must use the ValyrianTech ComfyUI template, which includes an nginx server for serving output images.

Docker images:

• valyriantech/comfyui-with-flux:latest - Includes FLUX.1 models pre-installed
• valyriantech/comfyui-without-flux:latest - Without FLUX models

Note: The template comes with FLUX.1 pre-installed, but GenSelfie requires FLUX.2. You will need to download the FLUX.2 models after deployment.

Note: FLUX.2 requires an RTX PRO 6000 GPU or equivalent.

Deploy on RunPod:

1. Get Hugging Face access:

   • Create a Hugging Face account if you don't have one
   • Go to FLUX.2-dev and request access to the gated repository
   • Create an access token at Hugging Face Settings

2. Deploy the ComfyUI pod:

   • Click this link to deploy: Deploy ComfyUI on RunPod
   • Select an RTX PRO 6000 GPU (or equivalent with sufficient VRAM)
   • Set your Hugging Face token in the pod's environment variables as HF_TOKEN

3. Download the models:

   • Once the pod is running, open a terminal in the pod
   • Run: bash /download_Flux2.sh
   • Wait for the models to download (this may take a while)

4. Get the ComfyUI URL:

   • In the RunPod dashboard, find the Direct TCP Port Mappings section (not the proxy URL)
   • Copy the IP and port (e.g., http://123.45.67.89:12345)
   • You'll need this for the COMFYUI_URL environment variable

Quick Start

1. Install dependencies:

   pip install -r requirements.txt

2. Configure environment:

   cp .env.example .env
   # Edit .env with your settings

3. Run the app:

   python main.py

   With verbose logging:

   python main.py --verbose

   All CLI options:

   python main.py --help
   # Options:
   #   -v, --verbose   Enable verbose logging
   #   --host HOST     Host to bind to (default: 0.0.0.0)
   #   --port PORT     Port to bind to (default: 8000)
   #   --reload        Enable auto-reload for development

4. Access the app:

   • Fan page: http://localhost:8000
   • Admin panel: http://localhost:8000/admin

5. First-time setup:

   • On first run, an admin password will be auto-generated and saved to .env
   • Check the console output or .env file for the password
   • Log in to the admin panel and configure your settings

Configuration

Environment Variables

Variable	Description	Required
ADMIN_PASSWORD	Admin panel password	Auto-generated
COMFYUI_URL	ComfyUI server URL	Yes
STRIPE_SECRET_KEY	Stripe secret key	For Stripe payments
STRIPE_PUBLISHABLE_KEY	Stripe publishable key	For Stripe payments
PUBLIC_URL	Public URL for Stripe redirects	For Stripe on RunPod
LNBITS_URL	LNbits instance URL	For Lightning payments
LNBITS_API_KEY	LNbits API key	For Lightning payments
DEBUG	Enable debug mode	No (default: false)
VERBOSE	Enable verbose logging	No (default: false)
DATA_DIR	Directory for database and uploads	No (default: /workspace)

The default DATA_DIR=/workspace works automatically with RunPod network volumes. For local development, set DATA_DIR=./data or another local path.

Stripe on RunPod

When running on RunPod with Stripe payments, you must set the PUBLIC_URL environment variable to your RunPod proxy URL. This is required because Stripe redirects users back to your app after payment, and it needs the public-facing URL (not the internal IP).

1. In the RunPod dashboard, find your pod's Proxy URL (e.g., https://abc123-8000.proxy.runpod.net)
2. Set this as the PUBLIC_URL environment variable or configure it in the admin panel under Stripe Settings

Without this, Stripe redirects will fail because they'll use the internal container URL which isn't accessible from the internet.

Project Structure

GenSelfie/
├── main.py                 # FastAPI entry point
├── config.py               # Settings & environment
├── database.py             # SQLite models
├── routers/
│   ├── admin.py            # Admin panel routes
│   └── public.py           # Fan-facing routes
├── services/
│   ├── comfyui.py          # ComfyUI API client
│   ├── social.py           # Social media profile fetching
│   ├── payments.py         # Stripe & LNbits integration
│   └── codes.py            # Promo code validation
├── templates/              # Jinja2 templates
├── static/                 # CSS, JS, uploads
├── workflows/              # ComfyUI workflow JSON files
├── input_examples/         # Example input images
└── genselfie.db           # SQLite database (created on first run)

RunPod Deployment

The easiest way to deploy GenSelfie is on RunPod:

1. Deploy GenSelfie: Deploy GenSelfie on RunPod
2. Select a CPU pod (no GPU needed for the web app itself)
3. Access the admin panel at your pod's proxy URL + /admin

See the ComfyUI Server section above for setting up the required ComfyUI backend.

⚠️ Important: If both GenSelfie and ComfyUI are running in the same RunPod datacenter, the connection may fail due to internal DNS issues. Deploy them in different datacenters to avoid this problem.

Docker Deployment

Pre-built Docker Image

The GenSelfie Docker image is available on Docker Hub:

docker pull valyriantech/genselfie:latest

Docker image: valyriantech/genselfie:latest

Running the Container

docker run -p 8000:8000 \
  -e ADMIN_PASSWORD=your_password \
  -e COMFYUI_URL=http://your-comfyui-server:8080 \
  -v ./data:/workspace \
  valyriantech/genselfie:latest

Environment variables can be passed with -e flags. See the Environment Variables section for all available options.

Persistent storage: Mount a volume to /workspace to persist the database and uploaded files. On RunPod, the /workspace directory is automatically mounted to the network volume.

Building Locally

To build the Docker image yourself:

# Build the image
docker build -t genselfie .

# Run locally
docker run -p 8000:8000 \
  -e ADMIN_PASSWORD=your_password \
  -e COMFYUI_URL=http://your-comfyui-server:8080 \
  -v ./data:/workspace \
  genselfie

Building and Publishing

A helper script is provided to build and push to Docker Hub:

# Build and push with today's date tag
python build_docker.py genselfie

# Build and push with custom tag
python build_docker.py genselfie --tag v1.0.0

# Also tag and push as :latest
python build_docker.py genselfie --latest

# Use a different Docker Hub username
python build_docker.py genselfie --username myusername --latest

Docker Image Details

The Dockerfile uses a multi-stage build:

1. Builder stage: Installs dependencies in a virtual environment
2. Production stage: Copies only the necessary files and virtual environment

The image:

• Based on python:3.12
• Exposes port 8000
• Uses /workspace as the default data directory (RunPod compatible)
• Runs via start.sh which starts the FastAPI server

Supported Social Platforms

• Twitter/X: Uses unavatar.io service
• Bluesky: Uses AT Protocol public API
• GitHub: Direct avatar URL
• Mastodon: Uses instance API with WebFinger
• Nostr: Uses nostrhttp.com API (supports npub, hex pubkey, and NIP-05 identifiers)

License

MIT