Telegram Academic Research Assistant Using Gemini AI and Decodo - n8n Workflow

• Researchers and academics needing rapid summaries of scholarly literature.


• Students who want to integrate AI research tools directly into their messaging apps.


• Developers looking for advanced n8n templates featuring Langchain agents and external tools.


• Automation specialists building complex chat-based research solutions using n8n.

This robust n8n workflow transforms a simple Telegram bot into a sophisticated, AI-driven academic research assistant. Leveraging the power of Gemini AI for complex reasoning and the specialized Decodo n8n node for targeted academic web scraping, this solution allows users to simply message their research query (via text, voice, or even an image of a document) and receive concise summaries of relevant academic papers. This n8n workflow is a prime example of high-value automation, combining conversational AI with data extraction to solve real-world research problems. The included flow control logic also ensures that if the output summary exceeds Telegram's message limit, the result is automatically sent as a downloadable text file, providing a seamless user experience within the n8n template structure. Using this n8n trigger and subsequent n8n nodes guarantees reliable research delivery.

The process begins with the Start Telegram Bot n8n trigger, waiting for an incoming message.

1. Input Analysis: The Detect Message Type n8n node determines if the input is text, an image, or a voice message.


2. Multimodal Processing: If a photo or voice message is received, it is downloaded via a Telegram n8n node. Gemini AI nodes (Analyze Image Content or Transcribe Voice Message) then convert the non-text input into actionable text data. This showcases the versatility of the n8n workflow when handling various inputs.


3. Agent Orchestration: The prepared data is passed to the Research Summary Agent n8n node. This specific n8n node is configured with detailed instructions (including insights about academic search URL parameters, often required in sophisticated n8n templates) and is equipped with the Simple Memory n8n node for conversational context and the specialized Decodo tool for web scraping.


4. Tool Use & Research: The agent analyzes the user's intent, generates targeted search queries (e.g., for Google Scholar or arXiv), and uses the specialized Decodo n8n tool to scrape relevant titles, abstracts, and publication details.


5. Summarization & Formatting: The agent summarizes the findings according to strict formatting rules.


6. Delivery Check: The Check Telegram Message Length n8n node validates the length of the final output. If the research summary is over 4000 characters, the n8n workflow automatically routes it through the Convert Output to Text File n8n node and sends it as a document. Otherwise, the Send Research Summary Message n8n node delivers the result directly as a text reply in Telegram. This complete n8n workflow showcases advanced flow control typical of useful n8n templates.

1. Import the n8n template: Copy the provided JSON and import it into your n8n instance.


2. Telegram Setup: Create a new bot using Telegram’s BotFather. Add your Telegram API credentials to the Start Telegram Bot n8n trigger and all subsequent Telegram n8n nodes.


3. Gemini AI Setup: Provide your Google Gemini API credentials to the Analyze Image Content, Transcribe Voice Message, Gemini URL Interpreter, and Gemini Research Model n8n nodes.


4. Decodo Setup: Obtain Decodo API credentials and configure the dedicated Decodo n8n node.


5. Configure Agent (Crucial Step): Run the isolated flow segment starting with Define Search URLs to generate the URL insights. Copy the generated Markdown text and replace the {{INPUTSEARCHURL_INSIGHTS}} placeholder in the Research Summary Agent's system message. This crucial step ensures the agent knows how to use the search links effectively within this n8n workflow.


6. Activation: Save and activate the n8n workflow to run in production.

Start Telegram Bot (n8n trigger): The primary n8n trigger for the entire process. It initiates the n8n workflow upon receiving any message update in the configured Telegram bot.
Detect Message Type (Switch n8n node): A critical flow control n8n node that routes incoming messages based on whether they contain text, a photo, or a voice message, ensuring the correct processing path is followed.
Decodo (Custom n8n node): Serves as the web scraping tool utilized by the AI Agent. It is specialized in extracting academic details (title, abstract, authors) from academic search engine results.
Research Summary Agent (Langchain Agent n8n node): The core intelligence of this n8n workflow. It uses the Gemini Research Model and the Decodo tool to execute a multi-step research plan, maintaining context via the Simple Memory n8n node.
Gemini URL Interpreter (Google Gemini Chat n8n node): Used during setup to analyze provided academic search URLs and automatically generate the necessary parameter documentation, thereby enhancing the agent's capability within this complex n8n template.
Check Telegram Message Length (If n8n node): An essential utility n8n node that manages output constraints. It checks if the AI-generated summary exceeds Telegram’s 4000-character limit, determining whether to send a direct message or convert the summary into a file.