AI Post Generation with Human Approval via Telegram - n8n Workflow

Social Media Managers: Seeking high-quality content at scale while maintaining human oversight.
Content Teams: Needing fast drafting capabilities coupled with an easy approval process.
Automation Enthusiasts: Looking for advanced examples of combining AI agents (Langchain) and communication tools within a single n8n workflow.
n8n Users: Interested in deploying complex Human-in-the-Loop (HIL) n8n templates.

Generating high-quality, brand-safe content automatically can be challenging. This advanced n8n workflow solves this by combining powerful AI generation with a mandatory Human-in-the-Loop (HIL) approval mechanism managed entirely through Telegram. The initial prompt kicks off the n8n workflow, which uses an AI agent powered by GPT 4.1 (via OpenRouter) and real-time search capabilities (Tavily) to create a compelling post draft. This draft is then sent to a human reviewer using a dedicated Telegram n8n node. Depending on the feedback, the post is either instantly published via the X n8n node or sent to a specialized Revision Agent for intelligent rewriting. This specific n8n template ensures speed and quality, proving the robust capabilities of n8n for modern content marketing.

This comprehensive n8n workflow starts with the Telegram_Trigger n8n trigger, which receives the initial command or topic.

1. Generation: The message is passed to the X Post Agent, an AI agent that uses the GPT 4.1 language model and the Tavily Search tool to gather information and generate a draft post for X (Twitter).


2. Preparation: The Set Post n8n node standardizes the data format.


3. Human Review: The Request Feedback Telegram n8n node sends the generated post draft to the designated channel or user for review, effectively implementing the HIL process.


4. Classification: The n8n workflow awaits a response. Once received, the Text Classifier n8n node uses a faster language model (2.0 Flash) to instantly classify the human feedback (e.g., 'Approve' or 'Revise').


5. Decision Branch:

To deploy this powerful n8n workflow, follow these steps:

1. Import: Copy the provided JSON data and import it into your n8n instance via the 'Workflows' section.


2. Telegram Setup: Configure the Telegram_Trigger and Request Feedback n8n node credentials. You will need a Telegram Bot token and the correct Chat ID for both the receiving feedback channel and the initial n8n trigger.


3. AI Service Setup: You must configure credentials for OpenRouter (used by the GPT 4.1 and 2.0 Flash Langchain n8n nodes) and Tavily Search (used as the real-time data retrieval tool).


4. X (Twitter) Setup: Configure the X Post n8n node with your X API credentials to allow for automated publishing.


5. Activation: Ensure all n8n nodes are correctly configured and activate the n8n trigger to run the workflow.

Telegram_Trigger (n8n Trigger): Starts the n8n workflow upon receiving a message, defining the initial input.
X Post Agent (Langchain Agent n8n Node): The primary generative AI engine. It drafts the X post content using GPT 4.1 and the Tavily Search tool for up-to-date context.
GPT 4.1 & 2.0 Flash (LM Chat OpenRouter n8n Nodes): Provides access to premium and fast large language models (LLMs) required by the AI agents for sophisticated reasoning and quick classification.
Tavily Search (Langchain Tool n8n Node): Used by the X Post Agent to perform web searches, ensuring the generated content is relevant and timely.
Request Feedback (Telegram n8n Node): Sends the draft post to the human reviewer for approval, crucial for the Human-in-the-Loop logic within this n8n workflow.
Text Classifier (Langchain Text Classifier n8n Node): Analyzes the human response received via Telegram, categorizing it as 'Approve' or 'Revise'.
Revision Agent (Langchain Agent n8n Node): An AI agent dedicated solely to refining the post based on specific human critique, maintaining high content standards.
X Post (Twitter n8n Node): The final action node, responsible for publishing the approved content to the X platform.