Content marketing automation tool

Landing Page 1
Landing Page 2
Choose a topic for your post
Pick a quote for image generation
Download your post

Inspiration

The project was inspired by the growing challenge content creators face: producing fresh, engaging social media content across multiple platforms while maintaining platform-specific best practices. We wanted a tool that not only generates content but truly understands the differences between LinkedIn’s professional tone, Twitter’s brevity, Facebook’s conversational style, and Instagram’s visual-first approach. The vision was to speed up content creation by stitching together multiple AI services into one cohesive, intelligent platform.

What it does

GenPostAI (fictional name) transforms a single topic idea into complete, ready-to-publish social media posts with matching AI-generated images.

It includes:

Platform-optimized text for LinkedIn, Facebook, X/Twitter, and Instagram
AI text generation via Llama 3.3 70B and DeepSeek
Intelligent quote extraction
AI image generation via Stable Diffusion 3.5 Large
Idea → Content → Image → Final Post progressive workflow
Supabase authentication
Built-in platform constraints

How we built it

We built the application using Kiro’s AI-assisted development workflow.

Vibe Coding -We described features in natural language and Kiro generated page structures, API integrations, and state management automatically.

Steering Docs

Inside .kiro/steering/ we defined:
tech.md — ES6 modules, API conventions, security
structure.md — 1:1:1 HTML/CSS/JS structure
product.md — platform rules, tone, length, user flow Architecture
Frontend: Vanilla JavaScript with ES6 modules
Backend: Express.js proxy for AI calls
AI Pipeline: Sequential processing with system prompts
State: localStorage + URL parameters
Authentication: Supabase

Challenges we ran into

API rate limiting and cost management
ES6/CommonJS module conflicts
Multi-page state persistence
Platform-specific quality tuning
Image/text coherence issues
Coordinating multiple AI services efficiently

Accomplishments that we're proud of

Full-stack AI integration (OpenRouter, Replicate, Supabase)
Production-ready code with secure proxying
Platform intelligence embedded into system prompts
Built entire application in ~8 hours using AI-assisted development
Clean architecture enforced through steering docs
True “Frankenstein spirit” combining text, images, authentication, and optimization

What we learned

How to orchestrate multiple AI systems efficiently How to encode engagement and platform rules for higher quality outputs How progressive UX dramatically improves completion rates How to migrate a full-stack codebase to ES6 modules How vibe coding + steering docs accelerate development Security best practices for modern AI apps

What's next for Content marketing automation tool

Carousel post generator for LinkedIn and Instagram Post scheduling, analytics, and A/B testing Custom prompt libraries and templates Multi-platform content repurposing MCP integrations (quotes, scheduling, LinkedIn previews) Advanced image settings and brand overlays Team collaboration and content calendars