ComfyUI Leon API Nodes

A custom ComfyUI node for generating images using Google's Imagen models via the HyprLab API.

Overview

This package provides a custom node called Leon_Google_Image_API_Node that integrates Google's Imagen models (imagen-4, imagen-3, imagen-3-fast) into ComfyUI workflows through the HyprLab API service.

Features

• Google Imagen Models: Support for imagen-4, imagen-3, and imagen-3-fast
• Multiple Aspect Ratios: 1:1, 3:4, 4:3, 9:16, 16:9
• Output Formats: PNG, JPEG, WebP
• Response Formats: URL or Base64 JSON
• Seed Support: For reproducible image generation
• Error Handling: Robust retry mechanism with exponential backoff
• Validation: Input validation including prompt length limits (10,000 characters)

Installation

1. Clone or copy this repository to your ComfyUI custom nodes directory:

   cd ComfyUI/custom_nodes/
   git clone <repository-url> comfyui-leon-nodes
   

2. Install the required dependencies:

   pip install -r requirements.txt
   

3. Restart ComfyUI

Usage

Node: Leon Google Image API

The main node for generating images with Google's Imagen models.

Inputs

Required:

• prompt (STRING): Main text input to guide image generation (max 10,000 characters)
• model (COMBO): Choose from imagen-4, imagen-3, or imagen-3-fast
• aspect_ratio (COMBO): Select from 1:1, 3:4, 4:3, 9:16, 16:9
• output_format (COMBO): Choose PNG, JPEG, or WebP
• seed (INT): Random seed for reproducible results
• api_url (STRING): HyprLab API base URL (default: https://api.hyprlab.io/v1)
• api_key (STRING): Your HyprLab API key

Optional:

• response_format (COMBO): Choose between "url" or "b64_json"

Outputs

• image (IMAGE): Generated image as ComfyUI tensor
• image_url (STRING): Image URL or base64 data URL
• seed (INT): The seed used for generation

API Configuration

To use this node, you need:

1. HyprLab API Key: Sign up at HyprLab to get your API key
2. API Access: The service provides access to Google's Imagen models with a 50% discount

Pricing (as per HyprLab documentation)

• imagen-4: $0.02 per image (50% off)
• imagen-3: $0.02 per image (50% off)
• imagen-3-fast: $0.01 per image (50% off)

Example Workflow

1. Add the "Leon Google Image API" node to your ComfyUI workflow
2. Enter your API key in the api_key field
3. Write your prompt in the prompt field
4. Select your preferred model, aspect ratio, and output format
5. Set a seed value if you want reproducible results
6. Connect the output image to a PreviewImage node or other image processing nodes

Error Handling

The node includes robust error handling:

• Automatic retry with exponential backoff (up to 5 attempts)
• Input validation (prompt length, empty prompts)
• API response validation
• Network error handling
• Detailed error messages for debugging

Technical Details

• Built using the same patterns as BillBum nodes for consistency
• Uses tenacity for retry logic
• Supports both URL and base64 response formats
• Automatically converts images to ComfyUI tensor format
• Includes proper image format conversion (RGBA)

Dependencies

• tenacity: For retry mechanism
• requests: For HTTP API calls
• pillow: For image processing
• torch: For tensor operations
• numpy: For array operations

License

This project follows the same license as the reference BillBum nodes.

Support

For issues related to:

• API access: Contact HyprLab support
• Node functionality: Create an issue in this repository
• ComfyUI integration: Check ComfyUI documentation

---

Based on the reference implementation from ComfyUI_BillBum_Nodes and integrated with HyprLab API.