How to Automate Research Report Generation with n8n

Research takes time. A lot of it. Between browsing Wikipedia, scanning Google Scholar, hunting for recent news, and compiling everything into a coherent document, hours disappear. What if you could type a single query and receive a polished, multi-source research report delivered straight to your inbox? That is exactly what this n8n workflow does.

This automation turns research from a marathon into a sprint. You provide a topic, and the workflow handles the rest: gathering data from academic papers, news articles, encyclopedias, and web sources, then packaging everything into a professionally formatted PDF. No copy-pasting. No manual formatting. Just results.

Prerequisites: What You Need to Get Started

Before building this workflow, gather your API credentials. This setup requires access to several services, each playing a specific role in the research pipeline:

• OpenAI API: Powers the query refinement process using GPT-4.1
• Google Gemini API: Drives the main research agent that coordinates all searches
• News API: Fetches recent news articles from 2024 to 2025
• Google Custom Search API: Enables general web searches
• SerpAPI: Provides access to Google Scholar for academic papers
• PDFShift API: Converts HTML templates into professional PDFs
• Gmail OAuth2: Sends the completed report via email
• Telegram Bot API: Delivers the PDF through Telegram

Key Components in n8n: This workflow uses 20 nodes including Manual Trigger, Code nodes for data processing, HTTP Request nodes for API calls, AI Agent nodes for intelligent research, and integration nodes for Gmail and Telegram distribution.

Step 1: Validate Input and Refine the Query

Every great research project starts with a clear question. The workflow begins with a manual trigger where you input your research query. A Code node validates that your query is at least three characters long, then cleans and timestamps it.

Here is where things get interesting. Instead of searching for just your original query, the workflow uses OpenAI GPT-4.1 to generate five related search queries. This expands your research scope automatically. Ask about artificial intelligence in healthcare, and the system generates queries covering diagnostics, treatment plans, ethical challenges, recent studies, and case studies. The Structured Output Parser ensures the AI returns data in a consistent JSON format that the workflow can process reliably.

Step 2: Deploy the Research Agent

The heart of this workflow is the Research AI Agent, powered by Google Gemini. Think of it as your personal research assistant that knows exactly where to look. This agent does not just search randomly. It follows a strategic approach, hitting four different sources simultaneously:

• Wikipedia: Grabs foundational knowledge and historical context
• Google Custom Search: Pulls general web information and recent developments
• News API: Finds current articles published between 2024 and 2025
• SerpAPI for Google Scholar: Retrieves academic papers from 2020 to 2025

image_1.png

Each search tool is configured as an HTTP Request node with specific parameters. The News API limits results to three articles, sorted by publication date. Google Search returns five results per query. Scholar searches focus on recent academic work. The Wikipedia API extracts plain text introductions without HTML markup.

What makes this powerful is the memory buffer. The AI Agent remembers context from previous searches within the same session, allowing it to build a comprehensive understanding of your topic across multiple sources.

Step 3: Process and Structure the Data

Raw research data comes back messy. The workflow uses a series of Code nodes to clean, parse, and organize everything into a structured format. First, the Parse Research Output node strips away any markdown formatting and converts the AI response into clean JSON.

The data then flows through a Split Out node that separates individual fields like introduction, summary, key findings, news highlights, scholarly insights, and sources. A Merge Split Items node recombines these elements into a cohesive structure, ensuring arrays are properly formatted and no data gets lost. An Aggregate node adds a timestamp and prepares the final dataset for PDF generation.

The workflow ensures every research report includes at least eight key findings, four news highlights, four scholarly insights, and eight unique sources. If initial searches fall short, the AI performs additional queries to meet these requirements.

Step 4: Generate a Professional PDF

Data without presentation is just noise. The Generate PDF HTML node transforms your research into a beautifully formatted document. This Code node builds an entire HTML template with:

• A gradient cover page featuring the topic and generation date
• Professional styling with custom fonts, colors, and spacing
• Highlighted sections for the first three key findings
• Linked sources for easy reference
• Page breaks and headers for print-ready output

The HTML includes CSS that creates a modern, clean aesthetic with a blue and white color scheme. Each section has proper hierarchy and breathing room. The design balances professionalism with readability.

image_2.png

Once the HTML is ready, an HTTP Request node sends it to PDFShift, which converts the template into a high-quality PDF. Another HTTP Request node downloads the completed file, making it available for distribution.

Step 5: Deliver the Report

A research report sitting on a server helps no one. This workflow delivers your PDF through two channels. The Gmail node sends a professionally formatted email with the report attached. The email template uses HTML styling to create an attractive message that introduces the report, explains its contents, and provides context for the recipient.

image_3.png

Simultaneously, a Telegram node sends the same PDF to a specified chat. This dual delivery ensures you get your research however you prefer to receive it, whether checking email at your desk or reviewing documents on mobile via Telegram.

image_4.png

Why This Workflow Changes Research

The benefits extend beyond saving time. This workflow creates consistency in research quality. Every report follows the same structure, includes the same types of sources, and maintains the same professional appearance. Teams can standardize their research processes without sacrificing depth or accuracy.

Use cases span industries. Academic researchers can quickly gather literature reviews. Marketing teams can compile competitive intelligence reports. Business analysts can track industry trends. Content creators can research article topics. Anyone who needs comprehensive information on demand benefits from this automation.

The workflow is also remarkably adaptable. Change the research sources by modifying the HTTP Request nodes. Adjust the report format by editing the HTML template. Switch delivery channels by replacing the Gmail and Telegram nodes with Slack, Discord, or any other platform n8n supports.

What once took hours now takes minutes. What once required manual formatting now happens automatically. What once lived in scattered browser tabs now arrives as a single, coherent document. That is the power of workflow automation done right.

Research will always require human insight and critical thinking. But gathering the raw materials? That is a job perfectly suited for automation. This n8n workflow proves that with the right tools and a bit of setup, you can transform tedious research tasks into elegant, automated systems that deliver real value.