Inspiration

While following YC startups and Hacker News, Mani and I noticed companies like Crustdata (live company data APIs) and Juicebox (AI-powered talent sourcing) gaining traction, with Crustdata’s founder being a friend’s husband. Seeing them raise funding and attract 200+ customers, we wondered: How hard would it be to replicate their core functionality but with a novel focus—"AI-powered startup scouting for investors?"

With Andreessen Horowitz as a hackathon sponsor and accelerator partner, we recognized that while a16z-backed startups unlock countless opportunities, investors still struggle to efficiently scout, assess, and connect with them. To solve this, we built StartupGPT in less than 8 hours, creating an AI-driven and Agentic voice-activated scouting platform where AI agents handle everything from NLP-powered search and ranking to real-time insights and automated outreach. Features AI-Powered Startup Search – Instantly find startups by simply describing what you're looking for. Intelligent Startup Insights – Get real-time data on founders, funding, traction, and risks. AI-Driven Investment Thesis Matching – Rank startups based on your unique investment criteria. Competitive Landscape Analysis – Compare startups with competitors using AI-driven insights. Automated Outreach & Follow-Ups – AI drafts personalized emails and tracks responses effortlessly. Conversational AI Interface – Simply talk to AI instead of manually searching databases.

What it does

StartupGPT is a AI-powered agentic scouting platform that helps VC's and angel investors discover, analyze, and engage with startups using a conversational AI interface. Every step of the process is powered by AI agents, from NLP-driven search to data aggregation, and even automating investor outreach. Using Eleven Labs AI voice tools, investors can simply ask, -“Find AI startups that raised over $1M”, and StartupGPT dynamically retrieves, ranks, and summarizes relevant startups in real time. AI agents handle everything query interpretation, data parsing, automated insights, and personalized follow-ups making startup scouting as seamless as a conversation. With Generative UI, investors get an intuitive, interactive experience, allowing them to explore opportunities effortlessly and find the next unicorn 🦄

Challenges We Ran Into

Real-Time Data Aggregation – Ensuring that StartupGPT fetches accurate and up-to-date information from multiple sources without lag using People Data Labs , Firecrawl and Perplexity AI. AI Agent Coordination – Seamlessly integrating AI agents and multi agent orchestration for search, analysis, and outreach while maintaining a conversational flow. Voice Activation & Responsiveness – Optimizing the voice interface to handle complex queries while keeping interactions natural. Scalability in Search & Matching – Designing an AI-powered ranking system that dynamically updates as new startups emerge. Building Under Time Constraints – We aimed to develop a fully functional product within just 8 hours at the hackathon, pushing ourselves to see how well we perform under pressure. UI Challenges – Ensuring smooth updates and dynamic management of tool responses as they varied across different websites and companies proved to be a complex task. LLM Models – Models like Perplexity and OpenAI exhibited emerging capabilities, and we sought to leverage those strengths effectively.

Accomplishments That We're Proud Of

Fully Functional working MVP in 8 Hours – From idea to execution, we built a working AI-driven scouting platform in record time. AI Agents Everywhere – Successfully automated even small tasks, ensuring investors get a seamless experience. Conversational AI for Investors – Making startup scouting as easy as talking to an AI assistant. Automated Outreach & Follow-ups – Investors can engage with founders instantly with personalized communication. Smart Insights & Red Flags Detection – AI not only finds startups but also highlights risks and opportunities dynamically.

What We Learned

AI Agents Unlock Efficiency – By using AI for every step, from discovery to outreach, investors can move faster than ever. Investors Need Real-Time Insights – Static databases aren’t enough—continuous monitoring of startups is essential. Conversational Interfaces Are Game-Changers – Talking to an AI is far more intuitive than using filters and manual searches. The Market Is Hungry for Better Scouting Tools – Seeing existing players like Crustdata and Juicebox succeed reinforced our belief in the demand for smarter scouting solutions.

What's Next for StartupGPT?

Let me answer in one sentence "Cover every startup and founder on the planet" :) Expand Data Sources – Integrate with platforms like Crunchbase, LinkedIn, and private deal flows to improve startup discovery. Enhance AI’s Decision-Making – Train our AI agents to predict investment potential based on market trends and founder credibility. More Personalization – Let investors set detailed investment theses and have AI refine recommendations accordingly. Investor Network & Collaboration – Allow VC's and angels to share insights and co-invest in promising startups. Scaling Beyond Hackathon – Turn this into a fully-fledged product and explore partnerships with accelerators like a16z and YC-backed startups.

Built With

  • nextjs
  • openai
  • resend
Share this project:

Updates

posted an update

Oops forgot to update individual contribution: @Mani led the frontend development, designing and implementing an intuitive user interface that ensures a seamless and efficient scouting experience for investors. I focused on backend integrations, setting up agent tooling, optimizing process, and configuring Eleven Lab's voice capabilities to enable a smooth conversational AI experience.

Log in or sign up for Devpost to join the conversation.