Multilingual YouTube Transcript Summarizer using Gemini and Google Docs - n8n Workflow

Content curators who need quick summaries of video content.
Social media managers preparing short-form content derived from long videos.
Knowledge workers who rely on n8n automation to process large volumes of information.
Users looking for advanced n8n templates integrating external APIs (RapidAPI) and large language models (Google Gemini).

This comprehensive n8n workflow solves the problem of manually transcribing and summarizing long YouTube videos. By leveraging a multi-step automation sequence, it converts a simple URL submission into a structured, multilingual summary saved directly to a Google Document.

Starting with an n8n trigger form submission, the system pulls the transcript via a third-party RapidAPI service. Crucially, the process utilizes custom scripting (n8n node) to clean and prepare the raw transcript data before feeding it into the Google Gemini LLM via the n8n AI Agent node. This ensures high-quality input for the summarization task. The final, optimized summary is then automatically appended to a designated Google Doc. This specific n8n workflow is a perfect example of how complex data processing and LLM interaction can be streamlined through robust n8n templates, saving significant time and resources.

This powerful n8n workflow follows a seven-step process:

1. Initiation (On form submission): The process begins with an n8n trigger, a built-in form, where the user submits the videoUrl and the desired output language (e.g., 'Spanish', 'French').


2. Data Mapping (Mapper): The input variables are securely mapped for use across the subsequent steps of the n8n workflow.


3. Transcript Retrieval (YouTube Transcript AI): An HTTP Request n8n node calls a RapidAPI endpoint, sending the YouTube URL to fetch the raw transcript data.


4. Data Preparation (Formator): A Custom Code n8n node executes JavaScript to parse the nested JSON response, decode Unicode characters, and clean the transcript, preparing it as optimal input for the AI model.


5. AI Summarization (AI Agent1): The prepared transcript is sent to the n8n AI Agent node, which uses the Google Gemini Chat Model. This n8n node is configured with specific system instructions to generate a concise, bulleted summary in the language specified by the initial form submission.


6. Output Extraction (Optimizer): Another Custom Code n8n node runs JavaScript to reliably extract only the summarized content, filtering out the surrounding prompt text and formatting applied by the AI model.


7. Documentation (Google Docs): The final, clean summary is inserted into a specified Google Doc using the Google Docs n8n node, completing the automation.

To deploy this advanced n8n workflow, follow these steps:

1. Import: Copy the provided n8n workflow JSON code and paste it into your n8n instance via the 'New' menu > 'Import from JSON'.


2. Credentials Setup:

1. Activation: Save and activate the n8n workflow to start processing video links automatically.

On form submission (n8n trigger): The starting n8n node. It generates a form to collect videoUrl and language inputs from the user.
YouTube Transcript AI (HTTP Request n8n node): Sends a POST request to the RapidAPI endpoint (youtube-transcriptor-pro.p.rapidapi.com) to fetch the transcript, parameterized by the videoUrl input.
Formator (Code n8n node): Essential for data hygiene. It contains custom JavaScript to decode Unicode escape sequences and handle complex JSON parsing of the raw transcript, preparing a clean string for the LLM.
AI Agent1 (n8n node): Orchestrates the summarization. It uses a Langchain agent structure, feeding the cleaned transcript and the required language into the Google Gemini Chat Model with a highly specific system prompt for structured output.
Optimizer (Code n8n node): Contains custom JavaScript logic (using regex) to ensure only the content designated under the 🎬 Summary heading is extracted from the verbose AI response, maximizing data clarity.
Google Docs (n8n node): The final action. Configured to 'Update' a specific document, inserting the clean summary output from the 'Optimizer' n8n node.