Frederik Takeru Sliwa@naganovitch630

Whom is a real-time market research engine that lets you talk directly to your customers — and your competitors’ customers — using nothing but a simple prompt. Instead of relying on surveys, outdated reports, or analyst teams, Whom pulls from what users are already saying across platforms like YouTube, GitHub, and Trustpilot. Our system scrapes public feedback at scale and filters out irrelevant noise — like “cool video!” or “nice voice” — using a fleet of modular, LLM-powered agents. What remains is only the most valuable insight: raw, unfiltered feedback revealing real pain points, feature gaps, unmet needs, and switching behavior. You can ask real strategic questions — “Why are users churning after onboarding?” “What makes RevenueCat better than our billing flow?” “What are early-stage founders asking for in 2025?” And get instant answers based on actual user experience — not assumptions, not secondhand interpretations. Under the hood, Whom is powered by a system of three autonomous agents: – A Manager Agent that processes your prompt and triggers scraping workflows – An Analyst Agent that filters and embeds meaningful feedback into a semantic vector database – A Chat Agent that retrieves contextually relevant insights and generates responses Tech Stack: Agent framework: Fetch.ai’s uAgents Scraping & hosting: Custom-built scrapers and agents deployed via Vultr Database & embeddings: Supabase with vector storage LLMs: LLaMA (8B & 70B) accelerated by Groq for real-time response Frontend: Built in Bolt.new with a desktop-native UX Whom was created in just three days for Raise Your Hack — but it’s designed for the future of how companies build, listen, and win.