Automated AI Research Paper Extraction & Summarization Database - n8n Workflow

1. Academic Researchers and Students: Those needing to stay updated on the latest scientific publications without manually reading full papers.


2. Data Scientists and AI Practitioners: Professionals who require a curated, summarized database of new AI and computer science research.


3. Automation Specialists: Users looking for advanced examples of chained AI processing, PDF handling, and reliable web extraction within an n8n workflow.


4. Anyone utilizing n8n templates for periodic data collection and analysis.

Keeping track of cutting-edge research can be overwhelming, especially in fast-moving fields like AI. This sophisticated n8n workflow solves the 'information overload' problem by automating the entire research digestion process. Using the specialized Decodo n8n node, the automation reliably pulls the newest AI paper listings from arXiv. It then leverages the power of Large Language Models (LLMs) via the OpenAI n8n node and LangChain integration to download the PDF, extract the content, and generate concise summaries. Finally, the results are meticulously cataloged in Google Sheets, creating a searchable research database. This specific n8n template demonstrates how to link scheduled triggers, complex data extraction, and powerful AI processing into a single, cohesive system.

This automation initiates its process using a schedule n8n trigger, configured to run every month, ensuring timely updates.

1. Data Acquisition: The Decodo n8n node is the first action, fetching the latest paper listings from the specified arXiv URL.


2. Content Extraction: The raw HTML output is processed by the Extract HTML Content n8n node, isolating the article listings. An OpenAI Chat Model and an Extract Articles LangChain n8n node then use structured prompting to reliably parse the title and PDF link from the raw text.


3. Iteration and Preparation: The results are prepared for batch processing using the Split Out and Loop Over Items n8n node. For each individual paper, the n8n workflow performs the following steps:


4. PDF Handling: The Get PDF n8n node (HTTP Request) downloads the research paper using the extracted link. The file is immediately passed to the PDF to Text n8n node, which performs optical character recognition (OCR) and extraction, converting the academic document into plain text.


5. AI Summarization Chain: The text is set into a field, then fed into the Paper Summarizer n8n node (LangChain). This node is configured with large chunk sizes to handle lengthy academic documents. It utilizes a dedicated OpenAI Chat Model n8n node (GPT-5-nano) to execute the summary generation.


6. Database Storage: The title, original URL, new summary, and extraction date are appended to a specified Google Sheet via the Store to database n8n node.


7. Notification: Once the summary processing cycle is complete, the Telegram Notifier n8n node sends an alert to a designated chat, confirming that the Google Sheets database has been updated with the latest summaries. This comprehensive n8n workflow exemplifies a complete data processing pipeline.

To deploy this expert n8n workflow, follow these steps:

1. Import the n8n Workflow: Copy the provided JSON data and paste it into your n8n instance using the 'Import from JSON' option.


2. Decodo Credentials: Set up your API key for the Decodo n8n node. This is crucial for reliable and rate-limited web scraping.


3. OpenAI Credentials: Connect your OpenAI (or compatible LLM) credentials. You need these for both the OpenAI Chat Model1 (information extraction) and the OpenAI Chat Model (summarization chain).


4. Google Sheets Setup: Create a new Google Sheet document. Authenticate the Google Sheets n8n node, select the correct document, and ensure your sheet has columns matching the required fields: 'url', 'title', 'summary', and 'extracted date'.


5. Telegram Configuration: Configure the Telegram Notifier n8n node with your Telegram Bot API key and the target Chat ID where you wish to receive notifications.


6. Review Schedule: Adjust the Schedule Trigger n8n node frequency if monthly updates are not suitable for your needs.


7. Activate: Once all credentials are set and the Google Sheets node is configured, activate the n8n workflow.

Schedule Trigger n8n Trigger: Starts the entire process automatically on a monthly basis, allowing for unattended operation of this n8n template.
Decodo n8n Node: Used for specialized, reliable fetching of web content from arXiv, bypassing potential issues related to site structure changes or rate limits. It fetches the latest 25 AI research listings.
Extract Articles (LangChain) n8n Node: Works in conjunction with GPT-5-mini to extract structured data (title and PDF URL) from the raw HTML content, transforming unstructured data into actionable items for the rest of the n8n workflow.
Get PDF (HTTP Request) n8n Node: Downloads the binary content of the research paper PDF based on the link extracted in the previous step.
PDF to Text n8n Node: Essential for AI processing. This converts the downloaded PDF file content into usable plain text.
Paper Summarizer (LangChain) n8n Node: This advanced summarization chain handles large documents by splitting them into chunks (Chunk Size 10000) and processing them sequentially to generate a full, high-quality summary. It requires the linked OpenAI Chat Model n8n node.
Store to database (Google Sheets) n8n Node: The final action, persisting the extracted research data, including the AI-generated summary, into a designated Google Sheets document.
Telegram Notifier n8n Node: Provides an instant communication update to the user upon the completion of the research update cycle.