Pitch Perfect

Inspiration

As startup founders, we've all been there - spending hours researching journalists, crafting personalized pitches, only to get crickets in return. Traditional PR agencies charge $5K+ monthly, and DIY approaches consume entire days with mediocre results. We realized AI could transform this broken process by automating the research, personalization, and timing that makes PR actually work.

What it does

Pitch Perfect is an AI-powered PR automation platform that turns weeks of manual work into a 5-minute daily routine. Users simply input their startup website, and the system: Analyzes your startup's content and positioning automatically Discovers trending articles from premium news sources using advanced search APIs Researches journalists' social profiles, writing styles, and recent focus areas Generates personalized Twitter and LinkedIn messages that reference specific articles Delivers 3 ready-to-send opportunities daily via email digest Each message feels human-written because it acknowledges the journalist's work, connects naturally to current trends, and positions your startup as a relevant expert source.

How we built it

Tech Stack: Frontend: Next.js 14 + TypeScript + Tailwind CSS + shadcn/ui Database: Supabase (PostgreSQL) APIs: Tavily (article discovery), Resend (email) Automation: Vercel cron jobs, Puppeteer web scraping AI: Custom message generation algorithms Architecture: Web scraping extracts startup positioning and keywords Daily cron jobs search Tavily API for journalist-written articles from premium sources Social media research finds journalist Twitter/LinkedIn profiles automatically Personalization engine generates tailored messages using article context and journalist background Email system delivers top 3 opportunities with ready-to-use messages

Challenges we ran into

Rate Limiting Hell: Balancing API calls across Tavily, social platforms, and web scraping without hitting limits required careful queue management and retry logic. Content Quality Control: Early versions generated generic messages. We had to build sophisticated relevance scoring and personalization algorithms to ensure each message felt genuinely human-written. Journalist Data Accuracy: Social media profiles change constantly. We implemented fallback strategies and data validation to maintain high-quality contact information.

Accomplishments that we're proud of

🎯 End-to-End Automation: Built a complete PR workflow that requires zero manual research - from article discovery to ready-to-send messages. 🤖 Human-Quality Personalization: Our AI generates messages that consistently pass the "human test" - they reference specific articles, acknowledge journalist expertise, and connect startup relevance naturally.

Built With

• nextjs
• supabase
• tavily
• vercel