🏆 Nike AI PDF Analyzer – Highlights from Nate Herk’s n8n Masterclass

Hey AI Skool Fam! 👟🤖

Super excited to showcase my latest project, built as part of Nate Herk’s awesome n8n Masterclass!

This journey brought me hands-on with vector databases, Google Gemini models, Pinecone, and the power of RAG (Retrieval Augmented Generation) for document Q&A. The star? An agent that reads Nike’s PDF reports and lets you pick its brain with natural language.

Ready for a whirlwind tour? Check out how it works—plus some flow screenshots to keep it visual!

💡 What Does This Project Do?

Uploads and analyzes Nike’s real PDF reports
Lets you chat with a bot that actually "read" the reports
Builds a powerful semantic search using Pinecone vector database and Google Gemini embeddings
Learns naturally over chat—remembers context, not just keywords!

⚙️ Workflow 1: Seamless Document Upload & Index

With a click on "Execute workflow," you upload the Nike PDF
The Default Data Loader grabs the whole file
The Recursive Character Text Splitter chops it into smaller, manageable sections (for smarter search)
Every chunk = embedding via Google Gemini
All embeddings indexed in Pinecone Vector Store for super-fast Q&A

⚡ Workflow 2: Chatbot Q&A Experience

User asks a question → Nike Agent (chatbot) gets the message
The agent uses Google Gemini Chat Model for smart responses
Simple Memory lets the bot remember our conversation flow
When you ask, Pinecone’s Vector Store finds the most relevant report snippet using embeddings from Google Gemini
The bot responds with info grounded IN the actual PDF, not just pre-trained fluff!

✨ Why This Project is Cool

Real, semantic Q&A: The bot finds relevant, context-rich answers beyond keywords—true document understanding!
Flows first, code minimalism: All built in n8n’s stunning visual interface. No tangled spaghetti code.
Vector DB skills unlocked: Pinecone + Gemini == blazing fast, meaningful document search (a must-know for future AI apps)
Context-aware conversations: It’s not just query/response—chats remember context from turn to turn.

🎓 A Huge Shoutout to Nate Herk & the n8n Masterclass

Every step of this build was inspired by techniques and best practices from Nate’s n8n Masterclass. If you’re thinking of diving deeper into no-code/low-code AI workflows, I highly recommend checking it out!

Looking forward to feedback 🚀

Tech Stack Highlights:

Happy building!

16 comments