Create a software agent (Python or Node.js) that completely replaces a human performing technical commands of ChatGPT within n8n.
What a person currently does manually —
the agent should do automatically, based on the text command from ChatGPT.
Agent Role Description
The agent is a technical executor who can:
create nodes,
change parameters,
insert code,
connect nodes,
turn on/off workflows,
run workflows,
analyze errors,
return reports.
ChatGPT gives a command → the agent does the work → the agent returns a response.
The user (human) does NOT give commands to the agent.
Interaction Scheme
ChatGPT ↔ Agent ↔ n8n
ChatGPT sends a text command
The agent interprets it
The agent performs the action via n8n API
The agent sends a report back to ChatGPT
🧰 Agent Functions
1. Working with n8n API
The agent must be able to:
2. Node Management
The agent must be able to do everything that a person does in the n8n editor:
create nodes (Function, Webhook, HTTP Request, Telegram, etc.)
set parameters (method, URL, headers, JSON body, raw body, expression fields, etc.)
insert and replace JavaScript code in Function nodes
connect nodes to each other
move nodes (position x,y)
delete nodes
The agent must be able to work with all data types used by n8n nodes:
JSON, raw body, headers, expression fields, HTTP request parameters, JavaScript code of Function nodes, etc.
3. Understanding Text Commands from ChatGPT
The agent must understand commands in free form:
“Create a Function node after Parse body”
“Insert this code into Format”
“Connect Parse → Format → Telegram”
“Replace method GET with POST”
“Check the last execution and tell me where the error is”
“Activate workflow 57”
The agent must:
interpret the text,
execute the command,
return a report.
4. Agent's Response to ChatGPT
The agent must return short, clear responses:
5. Communication Interface
The agent can have a simple interface:
Requirements:
What We Will Provide
n8n URL
API-key
examples of workflows
list of commands
test scenarios
What is Required from the Developer
Cost
Deadline
Short description of implementation
Demonstration of the agent's work (video or GIF)