StartupWiki is a community-driven research directory of global startup ventures. It's a living encyclopedia that uses a multi-agent AI system to research, verify, and compile detailed profiles of private companies — from early-stage startups to growth-stage scale-ups.
Unlike traditional databases that rely on manual curation or paid API feeds, StartupWiki combines crowdsourced suggestions with automated AI-powered research to create comprehensive, fact-checked profiles that include financial analysis, competitive positioning, SWOT analysis, and editorial spotlights.
Quickly evaluate startups with AI-generated financial models, unit economics, VC bull/bear debates, and market size analysis — all grounded in real web search data.
Discover emerging companies across categories like AI, Biotech, CleanTech, FinTech, and more. Each profile includes competitor mappings and differentiation analysis.
Browse the directory by category to understand which sectors are active, which startups are hiring, and where venture capital is flowing.
Anyone can suggest a startup for profiling. Submissions enter a review queue, are approved by editors, and are automatically researched and compiled into the directory.
Submit the name, industry, and optional pitch for any private startup. Your suggestion enters a moderation queue for editorial review.
Editors review the suggestion to ensure it meets the directory guidelines (private companies only, meaningful information provided).
Once approved, an autonomous AI pipeline fires off — SEC pre-clearance, deep web research, financial sandboxing, VC debate simulations, and editorial synthesis — all visible in real-time via live streaming.
The completed profile — featuring financial metrics, SWOT, competitive analysis, editorial spotlight, and sourcing — is added to the public directory for everyone to explore.
StartupWiki uses a multi-agent AI system orchestrated through a Node.js/Express server. Each agent performs a specialized role: SEC pre-clearance, factual research and verification, financial modeling, VC bull/bear debate, and editorial compilation. All research is grounded in live web search results and industry benchmarks.
The frontend is built with React, TypeScript, and Tailwind CSS. Live streaming of the generation process uses Server-Sent Events (SSE) for real-time updates.