Linky

• 

  Home Screen

• 

  Personalised Content Calendar

• 

  Customised Buckets and pillars

Inspiration

LinkedIn content creation is time-consuming and inconsistent. I built Linky to automate personalized, engaging LinkedIn posts using AI.

What it does

AI-powered LinkedIn automation platform with:

• Personalized content generation using multiple LLMs
• RAG system that learns from user interactions
• Automated scheduling with Redis queues
• Chrome extension for profile data extraction
• Real-time news integration for relevant content

How we built it

• Frontend: Next.js 15, TypeScript, Tailwind CSS
• Backend: Node.js, Express, PostgreSQL + pgvector
• AI: LangChain, OpenAI embeddings, multi-LLM support
• Infrastructure: Redis queues, AWS S3, Vercel deployment

Challenges we ran into

• LinkedIn API complexity and rate limits
• Making AI content sound human and personalized
• Real-time content relevance
• Scalable job processing for scheduled posts

Accomplishments that we're proud of

• Built comprehensive RAG system with vector similarity search
• Multi-LLM integration with intelligent fallback
• Production-ready queue system with monitoring
• Seamless user onboarding in under 5 minutes

What we learned

• User feedback loops are crucial for AI improvement
• Multiple LLM providers provide redundancy and cost optimization
• Vector databases excel at semantic search and recommendations
• Personalization significantly improves engagement

What's next for Linky

• Enhanced AI tools (hook generator, carousel maker)
• Multi platform support for all social media out there.
• Mobile app and team collaboration features
• Video content generation and voice cloning
• Multi-platform expansion beyond LinkedIn
• Enterprise white-label solutions

Built With

• anthropic
• apify
• bull
• claude
• gemini
• langchain
• next
• node.js
• openai
• postgresql
• react