AI-Powered Website Scam Risk Detector using GPT-4o and SerpAPI - n8n Workflow

Security analysts needing quick risk assessment tools.
E-commerce users validating new suppliers or questionable websites.
Individuals who frequently encounter suspicious links and need a reliable n8n workflow solution.
Anyone looking for advanced, multi-agent AI solutions utilizing complex n8n templates.

Analyzing a potentially fraudulent website manually is time-consuming, prone to oversight, and often inconclusive. This sophisticated n8n workflow solves this critical security challenge by deploying four specialized LangChain AI agents simultaneously. Each agent—focused on Technical Details, Search Signals, Pricing, and Content—uses GPT-4o and a SerpAPI n8n node to gather real-time data from the web. The results are then aggregated and fed to a final Evaluator agent, which synthesizes the data to provide a clear scam risk score from 1 (trustworthy) to 10 (highly suspicious). This powerful use of n8n templates dramatically accelerates due diligence, making it one of the most practical n8n automation tools for security checks.

This comprehensive n8n workflow starts with the On form submission n8n trigger, where the user inputs the target URL via a web form.

1. Parallel Analysis: The URL input from the n8n trigger is immediately dispatched to four independent AI agents. These agents (powered by GPT-4o and utilizing a dedicated SerpAPI n8n node for real-time searching) concurrently investigate different aspects: domain details (age, TLD), online scam reports, product pricing patterns, and website content quality.


2. Data Collection: The specialized reports generated by these four parallel processes are collected and organized using the Merge and Aggregate n8n nodes, ensuring all gathered evidence is compiled into a single data item.


3. Final Scoring: The complete dossier is passed to the final Evaluator n8n agent, which utilizes GPT-4o-mini. This crucial n8n node assesses all the red flags reported by the preliminary agents, assigns a definitive scam risk score (1-10), and presents the summarized findings, including a mandatory disclaimer. This complex orchestration demonstrates the power of the n8n automation platform to create expert AI systems.

To use this n8n workflow template effectively, follow these setup steps:

1. Import the Workflow: Copy the provided JSON code and paste it into your n8n instance using the "New -> Import from JSON" option.


2. OpenAI Credentials Setup: This n8n workflow requires access to the OpenAI API (specifically GPT-4o and GPT-4o-mini models). You must create a credential for your OpenAI API Key and connect it to all OpenAI Chat Model n8n nodes used in the agents and the final evaluator. Be aware that running this n8n automation may incur small usage costs (estimated around $0.01 per run).


3. SerpAPI Credentials Setup: Create an account on SerpAPI to access Google Search functionality. Obtain your API key and set up the corresponding credential in n8n. Connect this SerpAPI credential to all four SerpAPI n8n node tools within the specialized agents. The free tier of SerpAPI offers 100 searches per month; this powerful n8n workflow utilizes 5-15 searches per full run.


4. Execution: Once credentials are set up, activate the n8n workflow. Use the On form submission n8n trigger to input the URL you wish to analyze. The final results will be logged in the execution history of the Evaluator n8n node.

On form submission (n8n Trigger): The essential starting point of the n8n workflow, responsible for capturing the URL provided by the end-user.
Domain & Technical Details (LangChain Agent with GPT-4o): An expert n8n node that focuses on technical red flags, such as checking if the domain is less than 6 months old, examining suspicious TLDs (.xyz, .top), and confirming HTTPS usage. It uses a SerpAPI n8n node for external checks.
Search Engine Signals (LangChain Agent with GPT-4o): Queries public reports and search results using the SerpAPI n8n node to detect existing warnings, low ratings, or widespread scam queries across forums like Reddit or Trustpilot.
Product & Pricing Patterns (LangChain Agent with GPT-4o): Evaluates commercial deceit signals, such as exaggerated discounts or high-end products being sold at unrealistically low prices.
Content Analysis (LangChain Agent with GPT-4o): Scans for textual inconsistencies, including poor grammar, suspicious review patterns (duplicates, vague phrasing), and lack of legitimate company contact information.
Merge & Aggregate (Core Logic n8n Nodes): These nodes are crucial for collecting and structuring the results from the four parallel agent streams before the final decision phase.

• Evaluator (LangChain Agent with GPT-4o-mini): The final synthesis n8n node. It combines all inputs, assigns the conclusive scam risk score (1-10), and generates the final, comprehensive report for the user, completing the n8n workflow.