Marketio

Find what’s viral. Create content designed to win on every platform all with 1 click.

Comment

Inspiration

Marketing teams and indie founders struggle to keep up with fast-moving content and ad trends. What works on TikTok today might be irrelevant tomorrow, and translating organic trends into high-performing ads for platforms like Facebook or Shopify takes time, testing, and creative effort. We wanted to build a tool that automatically analyzes what’s actually performing right now and turns those insights into actionable, ready-to-use video and ad creatives tailored to a specific product and niche.

What it does

Marketio automatically discovers trending TikTok videos in a product’s niche, analyzes them using AI, extracts winning patterns, and generates both viral video concepts and ready-to-use scripts. It also generates AI-powered ads for platforms like Facebook and Shopify, including ad copy and creatives. For visual ad generation, we use Nano Banana to produce scroll-stopping ad creatives optimized for paid social and ecommerce use cases. Users can go from idea → insight → content or ads in minutes.

How we built it

First, we prompt the user to describe their product and niche. This allows us to accurately analyze the product and determine the most relevant niche it belongs to. That information is then sent to an n8n workflow, which kicks off the automation pipeline.

Using n8n, we automatically scrape top-performing TikTok videos from the past week within the user’s niche. The collected metadata is stored in an Airtable database. We then loop through each TikTok video, download it, upload it to Google Drive via the Google Drive API, and send the video to Gemini for in-depth analysis, including hooks, structure, creative patterns, and performance signals.

Once all videos are analyzed, we aggregate the results and send them to ChatGPT, which identifies overall trends, winning ideas, and recurring patterns. From this analysis, we generate:

Viral trend insights

High-converting hooks

UGC-style scripts for creators

A Veo 3 video generation prompt tailored to the user’s product

AI-generated ad creatives and copy for Facebook Ads and Shopify, powered by Nano Banana

All results are structured into a clean JSON response and sent back to the web app. From there, the web app displays the insights to the user. With one click, users can generate a custom AI video using Veo 3, fully aligned with their product, niche, and the viral trends we discovered. We also use Gemini 2.5 Flash to generate UGC-style scripts for users who want to create videos themselves in real life, and Nano Banana to generate scroll-stopping visual ad creatives optimized for paid social and ecommerce.

Accomplishments that we're proud of

We’re especially proud of building a system that scrapes and analyzes TikTok videos one by one to uncover real performance patterns, something we had never done before. Learning and using n8n from scratch was challenging, but incredibly rewarding once the workflow successfully produced clean, actionable data.

We’re also proud of the AI video generation feature. Using real trend insights, our system automatically generated marketing videos for a real product, which we then posted to brand-new TikTok, Instagram, and YouTube pages. In under 24 hours, each video surpassed 100 views on every platform, with one video reaching over 2,000 views. Seeing fully AI-generated content not only render correctly, but also perform organically in real marketing environments, felt like a major milestone and strong validation of our approach.

What we learned

How to design and debug complex n8n automation workflows

How to handle long-running async processes and API timeouts

How to structure AI outputs into production-ready JSON

How to extract meaningful creative insights from raw video data

What's next for Marketio

Next, we want to allow users to automatically publish generated videos to YouTube, Instagram, and TikTok directly from the platform with a single click. This will be handled through an n8n workflow that downloads generated videos from Google Drive and uploads them to each platform. Due to budget constraints and API access limitations, we weren’t able to obtain all required publishing credentials during the hackathon.

We also plan to expand our AI-powered ad generation capabilities beyond the current Facebook and Shopify support. This includes generating more creative variations, thumbnails, and platform-specific formats for Instagram Ads and other paid social channels, while continuing to leverage Nano Banana to produce high-performing visual ad creatives from real trend data.

Additionally, we want to analyze more videos per run and fully eliminate timeout issues. Improving batching, parallelization, and workflow efficiency will allow for stronger trend detection, deeper creative insights, and higher-quality video and ad outputs. Optimizing scalability and execution speed is our top technical priority moving forward.

Share this project:

Updates