Back to Templates

Batch Upscale Portrait Photos with Real-ESRGAN AI, Google Drive and Airtable

Created by

Created by: Robert Schröder || saits

Robert Schröder

Last update

Last update a month ago

Share


Portrait Photo Upscaler Workflow

Overview

Automated workflow that retrieves portrait photos from Airtable, upscales them using AI, and stores the enhanced images in Google Drive with organized folder structure.

Features

  • Automated Folder Creation: Creates timestamped folders in Google Drive
  • AI-Powered Upscaling: Uses Replicate's Real-ESRGAN for 2x image enhancement
  • Batch Processing: Handles multiple images efficiently with loop processing
  • Error Handling: Continues processing even if individual images fail
  • Airtable Integration: Retrieves images from specified database records

Prerequisites

Required Credentials

  • Google Drive OAuth2 API: For folder creation and file uploads
  • Airtable Token API: For accessing image records
  • Replicate HTTP Header Auth: For AI upscaling service

Airtable Setup

  • Column name: PortraitFotoAuswahl
  • Column type: Attachment field containing image files
  • Required: Valid Airtable Base ID and Table ID

Workflow Steps

  1. Manual Trigger: Initiates the workflow execution
  2. Create Folder: Generates new Google Drive folder with custom name
  3. Get Airtable Record: Retrieves specified record containing portrait images
  4. Extract URLs: Processes attachment URLs from Airtable field
  5. Loop Processing: Iterates through each image for individual processing
  6. AI Upscaling: Enhances images using Replicate's Real-ESRGAN (2x scale)
  7. Download Results: Retrieves processed images from Replicate
  8. Upload to Drive: Stores enhanced images in created folder

Configuration

Required Inputs

  • Folder Name: Custom name for Google Drive folder
  • Record ID: Airtable record identifier containing images
  • Base ID: (configurable)
  • Table ID: (configurable)

Upscaling Settings

  • Scale Factor: 2x (configurable)
  • Face Enhancement: Disabled (configurable)
  • Model: Real-ESRGAN v1.3

Technical Details

Node Configuration

  • Error Handling: Continue on individual failures
  • Response Format: File binary for image processing
  • Naming Convention: LoRa{timestamp}.png
  • Batch Processing: Automatic item splitting

API Endpoints

  • Replicate: https://api.replicate.com/v1/predictions
  • Model Version: nightmareai/real-esrgan:f121d640bd286e1fdc67f9799164c1d5be36ff74576ee11c803ae5b665dd46aa

Use Cases

  • Portrait photography enhancement
  • Batch image processing for portfolios
  • Automated content preparation workflows
  • Quality improvement for archived images

Output

  • Enhanced images stored in Google Drive
  • Organized folder structure with timestamps
  • Preserved original filenames with processed variants
  • Failed processes continue without stopping workflow

Template Benefits

  • Scalable: Processes unlimited images in single execution
  • Reliable: Built-in error handling and continuation logic
  • Organized: Automatic folder management and file naming
  • Professional: High-quality AI upscaling for commercial use