.jpg)
Process description
Automation is an intelligent Telegram bot for creating N8N workflow through natural language. The user describes the required workflow with text in a chat, after which an AI agent based on OpenAI GPT-5.1 and the full N8N documentation generates a valid JSON workflow, automatically creates it in the N8N system via the API and returns the user a direct link to the created workflow. The bot has a dialog memory for each user, which allows you to refine details and make changes to the workflow. The entire process from description to finished workflow takes 30-60 seconds, fully automating the creation of automations.
API keys and services
Telegram Bot API - receiving messages from a user
OpenAI API - GPT-5.1 for generating workflow JSON
Google Docs API - Read N8N documentation from Google Doc
N8N API - creating workflow programmatically through the API
System architecture by blocks
SECTION 1: GETTING A REQUEST FROM A USER
1.1 Telegram Trigger — Entry Point
Purpose: Receive text messages from a user describing the required workflow
Settings:
Type: N8n-Nodes-Base.TelegramTrigger
Version: 1.2
Updates: message - track new text messages
Credentials:
Telegram API
ID: SF7uluhdoba3dui5
Name: Workflow creator
The structure of the incoming data:
message.text - user request text with workflow description
message.chat.id - unique chat ID for personal memory and response
Examples of requests:
“Create a workflow that gets data from the API every day at 9am and saves it to Google Sheets”
“We need workflow to parse several pages with a delay between requests”
“Automate Slack notifications under certain conditions”
SECTION 2: DOWNLOADING N8N DOCUMENTATION
2.1 Get a document - Read documentation
Purpose: Obtaining up-to-date and complete documentation on creating N8N workflow from Google Docs
Settings:
Type: N8N-Nodes-Base.GoogleDocs
Version: 2
Operation: get
Credentials:
Type: Google Docs OAuth2 API
ID: Ne2xxS2MZ34C24OD
Name: Google Docs account 3
Document content:
This document contains full documentation for all popular N8N nodes:
- Basic workflow JSON structure
- Examples of all types of nodes with current parameters
- Expressions syntax for dynamic data
- Rules for creating connections between nodes
- Ready-made workflow templates for typical tasks
Result: The documentation text is sent to the AI Agent for use in generating the workflow
SECTION 3: AI WORKFLOW GENERATION
3.1 AI Agent - Workflow Generator
Purpose: An intelligent agent for analyzing user requests and generating a valid JSON workflow for N8N
Settings:
Type: @n8n /n8n-nodes-langchain.agent
Version: 3
Prompt Type: define
Text (Input):
= {{$ ('Telegram Trigger') .item.json.message.text}}. Here's detailed documentation for creating a workflow for n8n using the code: {{$json.content}}
Logic: Combines a user request with full documentation from Google Docs
System Message (short version, full prompt contains ~500 lines):
ROLE:
You're an expert at creating N8N workflow. Your task is to create a valid JSON workflow for N8N that is ready for import. ONLY JSON in the response, no comments or descriptions.
CRITICALLY IMPORTANT - WITHDRAWAL BANS:
It is ABSOLUTELY FORBIDDEN to reply with:
- Any descriptions of how to use the tools
- Any text up to JSON
- Any text after JSON
- Markdown code blocks
- Comments and explanations
THE ONLY ALLOWED OUTPUT IS:
{valid JSON workflow}
The first character of the answer is {
The last character of the answer is}
WORK ALGORITHM:
STEP 1: SILENT SEARCH
If you have access to search tools, use them SILENTLY
Find documentation for the N8N nodes you need
Examine the information found
STEP 2: CREATING JSON
Create a JSON workflow and immediately output it without prefaces
N8N DOCUMENTATION (included in the prompt):
The basic structure of the workflow is:
{
“name”: “Workflow Name”,
“nodes”: [{node parameters}],
“connections”: {node connections},
“settings”: {"ExecutionOrder”: “v1"}
}
Current node types (examples from documentation):
1. Manual Trigger - N8N-Nodes-Base.ManualTrigger v1
2. Schedule Trigger - N8N-Nodes-Base.ScheduleTrigger v1.2 with cron expressions
3. HTTP Request - N8N-Nodes-Base.HttpRequest v4.2
4. Code Node - n8n-nodes-base.code v2
5. Set Node - n8n-nodes-base.set v3.3
6. IF Node - n8n-nodes-base.if v2
7. Split In Batches - N8N-Nodes-Base.SplitinBatches v3
8. Wait Node - n8n-nodes-base.wait v1.1
9. Loop Node - n8n-nodes-base.loop v1
10. Merge Node - n8n-nodes-base.merge v2.1
11. Google Sheets - N8N-Nodes-Base.Googlesheets v4.5
12. Telegram - n8n-nodes-base.telegram v1.1
13. Slack - n8n-nodes-base.slack v2.1
14. Gmail - n8n-nodes-base.gmail v2.1
Expressions for dynamic data:
{{$JSON.fieldName}} - data access
{{$node ["Node Name"] .json.field}} - specific node data
{{$now}} - current date/time
{{$json.value > 100? “high”: “low”}} - conditions
WORKFLOW CREATION RULES:
1. Unique IDs for each node
2. Positions: start [250, 300], each next +200 by X
3. All data in the format [{"json”: {...}}]
4. Code Node always wraps the result in {json: {...}}
5. Credentials use your_API_KEY placeholders
READY-MADE TEMPLATES:
- Schedule+API+ Save
- Web scraping of multiple pages
- Filtering+Conditional processing
CHECKLIST BEFORE WITHDRAWAL:
- All IDs are unique
- All TypeVersion are correct
- All connections are configured
- Positions are logical
- JSON is valid without trailing commas
- NO text before/after JSON
- NO markdown blocks
3.2 OpenAI Chat Model1 - AI model
Purpose: Language model for generating workflow JSON
Settings:
Type: @n8n /N8N-Nodes-Langchain.lmchatopenai
Version: 1.3
Model:
Resource Locator Mode: list
Value: gpt-5.1
Cached Result Name: gpt-5.1
Credentials:
OpenAI API
ID: 7IAFMtudlwopFPA1
Name: OpenAI account 6
Built-in Tools: empty object - no built-in tools are used
Connection: Connected to AI Agent as AI_LanguageModel
Note: GPT-5.1 is used - the most advanced model for complex code generation tasks
3.3 Simple Memory - Dialogue memory
Purpose: Personal memory for each user to maintain the context of the conversation
Settings:
Type: @n8n /N8N-Nodes-Langchain.MemoryBufferWindow
Version: 1.3
Session ID Type: CustomKey
Session Key: = {{$ ('Telegram Trigger') .item.json.message.chat.id}}
Description:
Uses chat.id as a session key - this means that each user has their own isolated memory
Buffer Window Memory stores the last N messages in context
Allows the user to refine workflow details in several messages
Examples of memory usage:
User: Create a workflow to parse the site
Bot: [creates basic workflow]
User: Add a 2-second delay between requests
Bot: [modifies workflow based on the previous context]
Connection: Connected to AI Agent as ai_memory
SECTION 4: CREATING WORKFLOW IN N8N
4.1 Create a workflow - Creating via API
Purpose: Programmatically create a workflow in N8N via the API using generated JSON
Settings:
Type: n8n-nodes-base.n8n
Version: 1
Operation: create
Workflow Object: = {{$json.output}}
Logic: Takes an output from an AI Agent (generated by JSON workflow) and creates a workflow
Request Options: empty object - default settings are used
Credentials:
Type: N8N API
ID: M7byl2ts8ublzXOB
Name: n8n account
The result:
A new workflow is created in the N8N system
An object with full workflow information is returned:
- id: unique ID of the created workflow
- name: workflow title
- active: activation status (false by default)
- nodes: an array of nodes created
- connections: configured connections
SECTION 5: RESPONSE FORMATTING
5.1 Edit Fields — Creating a link
Purpose: Formatting a direct link to the created workflow for sending to the user
Settings:
Type: n8n-nodes-base.set
Version: 3.4
Assignments:
ID: 5fa01202-4554-4949-b5a2-81db9e20cd4e
Name: Link
Type: string
Value: = https://ваш-н8н.cloud/workflow/ {{$json.id}}
Logic:
Takes the ID of the created workflow from the Create a workflow node
Generates a direct link to workflow in the N8N interface
The user can immediately click on the link and start working with workflow
An example of the result:
https://ваш-н8н.cloud/workflow/abc123def456
Options: empty object
SECTION 6: SUBMITTING RESULTS
6.1 Send a text message - Reply to a user
Purpose: Sending a link to the created workflow to a Telegram user
Settings:
Type: n8n-nodes-base.telegram
Version: 1.2
Resource: message
Operation: SendMessage
Chat ID: = {{$ ('Telegram Trigger') .item.json.message.chat.id}}
Logic: Sends a response to the same chat where the request came from
Text: = {{$JSON.link}}
Logic: Sends a formatted link from Edit Fields
Additional Fields:
AppendAttribution: false - don't add a signature about the bot
Credentials:
Telegram API
ID: SF7uluhdoba3dui5
Name: Workflow creator
An example of a message to a user:
https://ваш-н8н.cloud/workflow/abc123def456
Required services and their configuration
Telegram Bot setup:
Creating a bot:
1. Open Telegram and search for @BotFather
2. Send /newbot command
3. Specify the bot's name (for example: N8N Workflow Creator)
4. Specify the bot's username (for example: n8n_workflow_creator_bot)
5. Get Bot Token
Setup in N8n:
1. Add Telegram API Credentials
2. Insert Bot Token
3. Webhook will be configured automatically
OpenAI API setup:
Obtaining an API key:
1. Go to platform.openai.com
2. Sign in to your account
3. Create an API key in the API Keys section
4. Copy key (shown only once)
Setup in N8n:
1. Add OpenAI Credentials
2. Insert API key
3. Select gpt-5.1 model in OpenAI Chat Model node
Note: GPT-5.1 requires sufficient balance and may have a high cost per token
Google Docs API setup:
OAuth2 setup:
1. Create a project in Google Cloud Console
2. Enable Google Docs API
3. Create OAuth2 credentials
4. Get Client ID and Client Secret
Setup in N8n:
1. Add Google Docs OAuth2 API credentials
2. Specify Client ID and Client Secret
3. Log in to OAuth
4. In the Get a document node, specify the ID of the document with the N8N documentation
Creating a document with documentation:
1. Create Google Doc
2. Insert full N8N node documentation from system message
3. Copy document ID from URL
4. Specify the ID in the Get a document node
N8N API setup:
Obtaining an API key:
1. In N8N, go to Settings - N8N API
2. Create a new API key
3. Copy key
Setup in N8n:
1. Add N8N API credentials
2. Specify Base URL (https://ваш-н8н.cloud)
3. Insert API key
Important: In the Edit Fields node, replace "https://ваш-н8н.cloud" with the real URL of your N8N instance
An example of workflow that automation has created:
Sending via WhatsApp, collecting data from Google sheets tables
Integrations and extensions
Possible additions:
Workflow library:
- Save all created workflows to the database
- Search for previously created workflows
- Reusing successful patterns
Versioning:
- Tracking changes in workflow
- Ability to roll back to a previous version
- Modification history
Team collaboration:
- Workflow sharing between users
- Comments and discussion
- Collaborative development
Analytics:
- Usage statistics
- Popular workflow types
- Success metrics
Testing:
- Automated testing of created workflows
- Verification of implementation
- Validation of results
Documentation:
- Autogeneration of documentation for workflow
- Description of each node
- Instructions for use
Practical value
For first-time N8N users:
- No need to know the N8N JSON structure
- No need to learn the interface
- You describe the task in simple language - you get a ready-made workflow
- Quick start with automation
For advanced users:
- Save time on routine workflow creation
- Rapid prototyping of ideas
- Create a workflow on the go via phone
- Use as a basis for further development
For teams:
- Centralized workflow creation
- Standardization of approaches
- Quick response to new tasks
- Democratization automation
For consultants and agencies:
- A quick demonstration of the N8N's capabilities
- Prototyping for customers
- Speeding up delivery projects
- Scaling up work