PassivePulse Finder

Finding profitable products for Marketers is time-consuming. PassivePulse Finder automates trend discovery, profit scoring, and campaign drafting while keeping humans in control.

Comment

Inspiration

Finding profitable products for affiliate marketing or Facebook ads is notoriously manual. Marketers spend hours scraping trends, guessing margins, and drafting campaigns—only to burn budget on unvalidated ideas. I wanted to automate the discovery and validation phase while keeping humans in control of every financial decision. PassivePulse was born from the idea that AI should handle the research, scoring, and drafting, but never spend money or publish without explicit approval.

What it does

PassivePulse Finder is a web app that helps users discover, score, and manage rising profitable opportunities for affiliate marketing and Facebook ads. Core features include: Trend Scanner: Pulls free Google Trends data (no API key) and surfaces rising product keywords Smart Scoring: Automatically estimates profit margin %, competition level, and projected ROAS Campaign Draft Generator: Creates ready-to-use Facebook ad copy, hooks, and UTM tracking structures. Approval Dashboard: Tracks opportunities through statuses (Discovered → Drafted → Approved/Rejected/Paused) with real-time analytics Human-in-the-Loop Oversight: Every status change requires explicit user confirmation—no auto-spend or auto-publish

How we built it

Platform: MeDo (AI-powered full-stack app builder) Workflow: Described the app in natural language → iterated via multi-turn chat → refined UI in visual editor → one-click deploy Frontend: React + Tailwind CSS (auto-generated by MeDo), fully responsive dashboard with filterable tables and status badges Backend/Storage: MeDo’s built-in serverless functions + integrated database storage. Originally designed for MongoDB Atlas, but pivoted to MeDo’s native storage to maintain functionality within free-tier constraints Data Logic: Free Google Trends via pytrends + CSV fallback, custom scoring algorithm for margin/competition/ROAS estimation Deployment: Published directly to a public URL via MeDo’s one-click deploy

Challenges we ran into

  • Backend connectivity: Initial MongoDB integration caused server crashes during MeDo’s deployment pipeline. Debugging was limited by token constraints and free-tier rate limits.
  • Storage pivot: Switched from MongoDB to MeDo’s built-in storage mid-build without breaking the approval workflow or UI state management.
  • Free-tier constraints: No paid APIs meant designing fallback data sources, mock scoring logic, and localStorage-style persistence for demo stability.
  • Human oversight balance: Ensuring the AI automated research and drafting while strictly enforcing approval gates required careful prompt engineering and state tracking.

Accomplishments that we're proud of

  • Built a fully functional, polished web app in under 48 hours using only AI chat and zero paid tools
  • Delivered a complete opportunity-to-approval workflow with real-time analytics and status tracking
  • Maintained strict human-in-the-loop design—critical for financial/marketing automation
  • Successfully navigated platform constraints by adapting storage architecture without sacrificing core functionality
  • Shipped a clean, mobile-responsive UI with professional UX patterns (filtering, badges, action buttons, analytics)

What we learned

  • AI app builders drastically reduce boilerplate but require clear, iterative prompting and fallback strategies
  • Human oversight is non-negotiable in automation—especially when dealing with budgets or live campaigns
  • Architectural flexibility (swapping databases/storage mid-build) can save a project when constraints hit
  • Free-tier development is viable if you design mock data, CSV fallbacks, and graceful degradation upfront
  • Prompt engineering for full-stack generation is becoming an essential modern development skill

What's next for PassivePulse Finder - Built with MeDo

  • Reintegrate MongoDB/Supabase with proper environment variables, connection pooling, and error handling
  • Add live API integrations (Meta Ad Library, Amazon Movers & Shakers, affiliate network feeds)
  • Implement performance tracking: auto-monitor ROAS, flag underperformers, and suggest budget reallocation
  • Add multi-user/team roles, campaign export (CSV/Ads Manager compatible), and scheduling features

Built With

  • googletrendapi
  • react
  • supabase
  • tailwind
Share this project:

Updates