Groq

📢 Winner announcement: Last month we hosted the AI Supaheroes Hackathon in collaboration with Supabase and AgentOPS In this hackathon, builders were asked to leverage one of our sponsors and build: 🦸♀️ AI 'supaheroes' and powers: empowering humans through AI We are excited to share our people's choice winner... Luna AI by Abdul Shahir 🎉 Luna AI is transforming the way educators and professionals create learning material. Inspired by the challenges faced in tailoring course material for specific groups, Luna AI was built to simplify and accelerate the course creation process. Whether you’re an educator looking to customize lessons or an expert aiming to monetize your knowledge, Luna AI has you covered. Luna has already created over 200 lessons in the last couple of weeks 🚀 Stack: - Groq - Supabase - React - Typescript Read more about Luna AI in the article linked in the comments👇 #hackathon #AI #EdTech #Elearning #builders

🤖Mixture-of-Agents Demo Powered by Groq This Streamlit application showcases the Mixture of Agents architecture proposed by Together AI, powered by Groq LLMs It allows users to interact with a configurable multi-agent system https://lnkd.in/guht9_eX

Project Completion: SearchGPT is Here !!! A week late, but here it is! The long, winding road of scraping, summarizing, and RAG-ing (yes, that's a verb now) is finally done. After countless hours of wrestling with scraping Google search results, taming websites, and figuring out how to cram context into LLMs (spoiler: you can’t just brute force it), I’m thrilled to say that SearchGPT is officially complete! 🎉 What’s the Final Result? A slick Streamlit-based app that: Scrapes Google Search Results (Yes, I’m talking Google Custom Search Engine without using Google Custom Engine 🕵️♂️) Extracts content from the websites Summarizes it all using the power of FAISS + LLMs + Retrieval-Augmented Generation (RAG). 💡 But Wait, There’s More: It’s a zero-cost project! 💰 I kept my promise of using only free tools—no payment details needed. Built on top of Groq's lightning-fast LLM service (seriously, these folks are generous). Everything is tied together in a neat Streamlit interface. Easy peasy. What Did I Learn? The first thing I learnt was being patient. RAG applications take time to load on mid-tier laptops, and now the technical part. Context is EVERYTHING. LLMs are very needy—just try to give them a chunk of data without context, and they’ll throw a tantrum. Vector DBs are a must. Hello, FAISS from Meta! Couldn’t have done it without you. Free tools can be powerful, but they require some extra love (and patience). What’s Next? Polishing the summarization process. Streamlining the UI for even better user experience. Maybe even an expansion to focus on the top 3 websites for more targeted summaries. Who knows, lol? Final Thoughts: This was no rocket science, but the journey of building SearchGPT taught me a ton about scraping, LLMs, and dealing with a million “minor” issues that quickly snowball into “major” ones. Feel free to check out the README for more details, and as always, I’m open to feedback and suggestions. Onward and upward! The only thing left is hosting this application. The only solution to make everything streamlined and crack proof (well there's nothing crack proof, but...) is getting this on Google Cloud Run service. Only there, I will require payment details. Do checkout the repo - https://lnkd.in/duTfpB-f , I do have attached the demo for the application. #AI #MachineLearning #SearchGPT #RAG #OpenSource #WebScraping #LLM

Manaflow (YC S24) is a simple way to automate repetitive office work in tables. Using natural language, you can program fleets of AI agents to execute millions of tasks involving data retrieval, web research, API integrations, and software actions. Today, we're launching a demo of our new interface! You can now program AI agents using a Notion-like user interface, create & call custom tools, populate Manasheet columns, and execute your repetitive workflows on a cron schedule. If you have repetitive tasks that you'd love to tackle with AI automation, please schedule a time with us here: https://lnkd.in/gXd2iMYG.