# Automatically generate visual presentations with n8n, GPT-5, and Gamma Source: https://www.youtube.com/watch?v=-nbHJzLpt18 Recap page: https://rapidrecap.app/video/-nbHJzLpt18 Generated: 2025-11-10T16:10:36.532+00:00 --- ## Quick Overview This demonstration outlines a full-stack content research, writing, and designing automation workflow using n8n, GPT-5, and Gamma to create visually engaging LinkedIn Carousels automatically based on user input for topic and audience. The workflow executes sequential steps including web research via Tavily, data aggregation, narrative drafting via a Storyteller Agent, content creation via Gamma's API, and final compilation into a referenceable format saved to a spreadsheet. **Key Points:** - The entire process is orchestrated through an n8n workflow designed for full-stack content automation, covering research, writing, and design. - The workflow begins by taking the topic ("AI Agent Research Insights and Findings an archive for October 2025") and audience ("AI Devs") from an initial form submission. - Web research is performed using the Tavily tool, followed by splitting, editing, and aggregating the retrieved information. - A Storyteller Agent, powered by GPT-5, drafts a cohesive narrative based on the aggregated research, adhering to a specific instruction set for style (e.g., LinkedIn Carousel style). - The drafted narrative is passed to the 'Create Content with Gamma' node, which uses Gamma's API (requiring an API key) to generate a presentation with visuals and slides. - After content generation, the workflow retrieves a URL for the presentation, waits for completion, and finally appends the Gamma URL and reference information to a Google Sheet for final compilation and reference. - The resulting Gamma presentation is highly visual, featuring content like 'October's AI Agent Revolution' and 'Evaluations Finally Grew Up' slides, demonstrating the automated design capability. ![Screenshot at 0:21: The visualization of the full n8n workflow titled 'Full-stack Content Research, Writing & Designing Automation' showing sequential nodes from input to final compilation.](https://ss.rapidrecap.app/screens/-nbHJzLpt18/00-00-21.png) **Context:** The video serves as a tutorial demonstrating how to build an advanced automation pipeline using the workflow automation tool n8n to fully automate the creation of rich media content, specifically LinkedIn Carousels. This process integrates several cutting-edge AI capabilities, notably leveraging the advanced language model GPT-5 for narrative structuring and the presentation tool Gamma for visual output, all tied together with custom logic and external services like Tavily for research and Google Sheets for output logging. ## Detailed Analysis The video details the construction and execution of an n8n workflow aimed at automating the creation of visually rich LinkedIn Carousels. The process starts with a form input defining the topic (e.g., "AI Agent Research Insights and Findings an archive for October 2025") and the target audience ("AI Devs"). The workflow proceeds through Web Research using Tavily to gather data, which is then processed (split, edited, aggregated). Next, a 'Storyteller Agent' node, configured to use an OpenAI Chat Model (implied GPT-5 capabilities), crafts a narrative based on the research, following specific style instructions like using a 10-point LinkedIn style recap. This narrative is fed into the 'Create Content with Gamma' node, which makes an API call to Gamma's 'generateContent' endpoint, passing the narrative, desired format ('social'), and style ('basic' or 'modern') via JSON body. The system then waits for the content to be generated, retrieves the resulting URL, and finally logs all generated assets, including the Gamma URL, into a Google Sheet under the 'Compile for reference' stage. The demonstration concludes by showing the generated Gamma presentation, which is visually structured with slides on topics like 'Evaluations Finally Grew Up' and 'Planning Gets Pragmatic,' confirming the successful integration of research, narrative generation, and visual design. ### Workflow Setup & Input - Enter trending topic and audience - Web Research via Tavily (includes steps: Develop Prompts, Split Out, Edit Format, Aggregate) - Draft Narrative via Storyteller Agent (using OpenAI Chat Model) ### Content Creation with Gamma - 'Create Content (Gamma)' node uses POST request with API key to generate content - 'Wait' node pauses execution for generation - 'Get URL' node retrieves the final presentation link ### Final Compilation & Output - 'Append to a sheet' node logs the Gamma URL and reference data to a Google Sheet - Final presentation showcases visually rich slides on AI agent topics like 'Evaluations Finally Grew Up' and 'Planning Gets Pragmatic' ### Storyteller Agent Configuration - Prompt instructs GPT-5 to create a cohesive narrative combining insights from all sources, using a 10-point LinkedIn style review, and adhering to specified tone/style parameters (e.g., conversational, forward-thinking). ![Screenshot at 0:00: The main n8n canvas displaying the full workflow pipeline for content automation.](https://ss.rapidrecap.app/screens/-nbHJzLpt18/00-00-00.png) ![Screenshot at 0:19: The initial input section where the topic and audience for the LinkedIn Carousel are defined.](https://ss.rapidrecap.app/screens/-nbHJzLpt18/00-00-19.png) ![Screenshot at 0:36: The Web Research stage of the workflow, highlighting the use of the Tavily tool for data gathering.](https://ss.rapidrecap.app/screens/-nbHJzLpt18/00-00-36.png) ![Screenshot at 0:41: The Draft Narrative stage, showing the Storyteller Agent connecting to an OpenAI Chat Model.](https://ss.rapidrecap.app/screens/-nbHJzLpt18/00-00-41.png) ![Screenshot at 2:21: The configuration panel for the Storyteller Agent, showing the detailed instructions provided to the LLM for narrative creation.](https://ss.rapidrecap.app/screens/-nbHJzLpt18/00-02-21.png) ![Screenshot at 3:15: The 'Create Content \(Gamma\)' node configuration showing the POST request parameters, including the Gamma API endpoint and header authentication.](https://ss.rapidrecap.app/screens/-nbHJzLpt18/00-03-15.png) ![Screenshot at 3:52: The output JSON from the 'Get URL' node, confirming successful content generation and providing the 'gammaUrl'.](https://ss.rapidrecap.app/screens/-nbHJzLpt18/00-03-52.png) ![Screenshot at 5:07: The final output logged in the Google Sheet, showing the generated URL and status updates.](https://ss.rapidrecap.app/screens/-nbHJzLpt18/00-05-07.png)