.jpg)
Process description
This automation is a full-fledged AI-powered photo editor that works via the Telegram interface. The system takes photos and text commands from users, uses an intelligent AI agent to understand queries, performs various operations with images (editing, merging, searching), integrates with Google Drive to store files, and provides a personalized experience through dialogue memory.
API keys and services:
- Telegram Bot API - for the user interface (bot: Photo editing 2)
- OpenAI API Key - for AI agent (gpt-4.1-mini)
- Google Gemini 2.5 Flash API - for editing images
- Google Drive OAuth2 - for storing and searching files (account 13)
- Sub-workflows - Edit Image and Combine Image for specific operations
System architecture by blocks
SECTION 1: MESSAGE PROCESSING
1.1 Telegram Trigger — Monitoring messages
Purpose: Tracks all incoming messages from users
1.2 Switch - Message type classification
Purpose: Separates messages into photos and text for different processing
Switch conditions:
- Output 0 “Photo”: $json.message.photo exists (there is a photo)
- Output 1 “Text”: $json.message.text exists (there is text)
SECTION 2: PHOTO PROCESSING
2.1 Download File - Download photos from Telegram
Purpose: Downloads a photo submitted by the user
Settings:
- Resource: file
- File ID: {{$json.message.photo [3] .file_id}} (high res)
- Credentials: Photo editing 2
Why index [3]: Telegram provides several photo sizes; the index [3] usually corresponds to the highest resolution.
2.2 Upload file - Save to Google Drive
Purpose: Automatically saves uploaded photos to Google Drive
Settings:
- Name: {{$now.format ('yyyy-mm-dd')}} (date name)
- Drive ID: My Drive
- Folder ID: your folder id
- Credentials: your credentials
2.3 Set Text - Preparing the context
Purpose: Generates a message for the AI agent with information about the downloaded file
Message text:
A user just sent a photo. Here's the Google Drive ID for this file: {{$json.id}}
SECTION 3: AI AGENT - SYSTEM CORE
3.1 Photoshop Agent — Main AI Assistant
Purpose: An intelligent agent that understands user requests and performs operations with images
Connected components:
- OpenAI Chat Model (gpt-4.1-mini) - main language model
- Simple Memory - memory of conversations with each user
- Change Name - file name change tool
- Combine Images - image merging tool
- Search Raw Files - search for files uploaded by the user
- Search AI Images - search for AI-generated images
- Edit Image - image editing tool
System prompt:
# Review
You're a personal assistant. Your job is to use the tools available to you to help the user fulfill their request.
## Tools
Change Name — use this tool to change the photo's name. To change the name, you must know the file ID.
Combine Images — use this tool to combine two images.
Search Raw Files — Use this tool to find image files that a user has previously uploaded.
Search AI Images — use this tool to search for AI-generated images that you helped the user create.
Edit Image — use this tool to edit an image.
## Instructions
1) If a user sends a photo, ask them what to call that photo by saying, “How do you want me to name this photo on your Google Drive?” Then, when he answers, change the name using the “Change Name” tool.
3.2 Simple Memory - Personalized memory
Purpose: Saves the context of the conversation for each user separately
Settings:
- Session ID Type: CustomKey
- Session Key: {{$ ('Telegram Trigger') .item.json.message.chat.id}} (unique chat ID)
- Context Window Length: 10 (last 10 posts)
Benefits: Each user has their own story, AI remembers previous operations and can link to previously processed images.
SECTION 4: AI AGENT TOOLS
4.1 Change Name - Rename files
Purpose: Allows AI to change file names in Google Drive
Google Drive Tool settings:
- Operation: update
- File ID: {{$fromai ('File_to_update')}} (receives from AI)
- New Updated File Name: {{$fromAI ('New_Updated_File_Name')}} (receives from AI)
4.2 Search Raw Files - Search for user files
Purpose: Searches for files that the user has previously uploaded
Settings:
- Resource: FileFolder
- Return All: true
- Folder ID: your folder id
- What To Search: files
- Options Fields: [id, name, WebViewLink]
4.3 Search AI Images — Search for AI-generated images
Purpose: Searches for images created with AI tools
Settings: Similar to Search Raw Files, but with a focus on AI-generated content.
4.4 Edit Image — Editing images
Purpose: Calls sub-workflow to edit images
Input parameters for sub-workflow:
- ImageTitle: The name of the new image
- ImagePrompt: Editing instructions
- ImageID: ID of the image being edited
4.5 Combine Images - Combine images
Purpose: Calls sub-workflow to combine two images
Input parameters for sub-workflow:
- prompt: Merging instructions
- image1: First image ID
- image2: Second image ID
- ImageTitle: Result name
SECTION 5: SUB-WORKFLOWS
5.1 Edit Image Sub-Workflow
The process:
The principle is the same as in Photo AD System automation
- When Executed by Another Workflow - gets parameters
- Download file - downloads the original image from Google Drive
- Extract from File - converts to Base64
- generate_image - edits via Gemini 2.5 Flash
- Convert to File - converts the result back
- Upload file - saves to Google Drive
- Result - returns a link to the result
Prompt for Gemini:
Edit the photo by this instruction: {{ImagePrompt}}.
The edited image must be in high quality!
And do not make any comments just send the edited photo.
5.2 Combine Image Sub-Workflow
The process:
The principle is the same as in the tool before
- Edit Fields - creates an array of two image IDs
- Split Out - separates for individual processing
- Download file - downloads both images
- Extract from File - converts to Base64
- Aggregate - combines data for a single request
- generate_image - combines via Gemini with two images
- Convert to File + Upload file - saves the result
SECTION 6: ANSWERS TO THE USER
6.1 Respond - Sending Responses
Purpose: Sends the AI agent's responses back to the user
Settings:
- Chat ID: {{$ ('Telegram Trigger') .item.json.message.chat.id}}
- Text: {{$json.output}} (response from an AI agent)
- Additional Fields:
- AppendAttribution: false
- parse_mode: HTML (formatting support)
Node connection diagram
Main stream:
- Telegram Trigger → Switch
Photo stream:
- Switch (Photo) → Download File → Upload file → Set Text → Photoshop Agent
Text flow:
- Switch (Text) → Photoshop Agent
The final answer is:
- Photoshop Agent → Respond
AI connections:
- OpenAI Chat Model + Simple Memory → Photoshop Agent
- Change Name + Search Raw Files + Search AI Images + Edit Image + Combine Images → Photoshop Agent
Required services and their settings
Telegram Bot setup:
- Create a bot via @BotFather
- Get the token for “Photo editing 2"
- Set up rights to receive photos and messages
Setting up Google Drive:
- Create a “Photos to automate” folder
- Folder ID: 12rfff5hacwvsde2nlcyzw7c2xje8avuj
- Set up OAuth2 for Google Drive account 13
- Permissions to read, write and modify files
Setting up AI services:
- OpenAI API for gpt-4.1-mini (main agent)
- Google Gemini 2.5 Flash for image editing
- Setting up billing for both services
Creating Sub-Workflows:
- Edit Image
- Combine Image
- Set up Execute Workflow Trigger for everyone
System capabilities
Image operations:
- Auto save photos in Google Drive
- Editing images by text instructions
- Combine two images into one
- Rename files at the user's request
- Search for previously downloaded files
AI capabilities:
- Understanding natural language - teams in Russian
- Contextual memory - remembers previous operations
- Intelligent tool selection - determines the required operation by itself
- Personalization - separate memory for each user
Integrations:
- Telegram interface - easy to use via messenger
- Google Drive storage - secure cloud storage
- AI processing - modern editing technologies
- Modular architecture - easy to add new features
Use cases
Scenario 1: Uploading and naming photos
👤 User: [sends photo]
🤖 AI: How do you want me to name this photo on your Google Drive?
👤 User: “Vacation in Paris”
🤖 AI: Done! Photo saved as “Vacation in Paris”
Scenario 2: Editing an image
👤 User: Change the sky color to sunset in my last photo
🤖 AI: [searches for the latest photo, edits via Gemini]
🤖 AI: Image edited! Here's a link: [Google Drive link]
Scenario 3: Combining images
👤 User: Combine “Vacation in Paris” and “Eiffel Tower” photos
🤖 AI: [finds both images, combines]
🤖 AI: Images are combined! A new “Paris collage” image has been created
System application
For personal use:
- Quick editing photos directly on Telegram
- Organizing a photo archive with smart names
- Creative experiments with image merging
- Mobile access to advanced AI tools
For small businesses:
- Product photo processing for an online store
- Creating collages for social media
- Fast retouching without Photoshop skills
- Teamwork via shared Google Drive
For content creators:
- Content processing on the go
- Experimenting with visuals for social media
- Creating unique images via AI
- Archiving works with a convenient organization
The result of the system
What happens is:
- Pocket AI photo editor voice-controlled
- Intelligent cloud storage images
- Personalized assistant with dialogue memory
- Modular system to expand functions
- Simple interface for complex operations
Performance metrics:
- Processing time: 30-60 seconds per surgery
- Editing quality: professional AI quality
- Usability: messenger interface
- Personalization: individual memory for each user
Advantages over traditional editors:
- Accessibility - works on any Telegram device
- Simplicity - natural language commands instead of a complex interface
- AI help - understands intentions without precise instructions
- Cloudiness - all files are automatically saved and synced
- memento - remembers previous operations and can refer to them
ROI and practical value:
- Save on software - replacement of expensive graphic editors
- Workflow acceleration - operations in seconds instead of minutes
- Lowering the entry threshold - you don't need graphics skills
- Mobility - full editing on a smartphone
This system turns Telegram into a powerful AI-powered photo editor with a personal assistant and cloud storage!