Inspiration

Every day, tech headlines are filled with new breakthroughs — AI models, platform launches, funding rounds — but people rarely translate these trends into actionable startup ideas. I built this bot to close that gap: delivering daily startup ideas based on real trending tech news.

What I Learned

  • Using News API effectively for gathering structured, trending tech news
  • Designing LLM prompts for summarization and creative idea generation
  • Building automation flows for Slack, Telegram, and email integration
  • Creating a reliable, daily content pipeline that combines APIs and AI

How I Built It

  1. News API integration – I fetch the latest tech/startup headlines and articles using News API, filtering by source, keyword, or category.
  2. Summarization + Idea Generation – I send the cleaned headlines or article descriptions to an LLM with a custom prompt to produce one-liner startup ideas.

Challenges Faced

  • Filtering out low-signal or repetitive news in API results
  • Tuning prompts so the LLM ideas were fresh, not generic
  • Managing API quotas (News API + LLM tokens)
  • Scheduling reliable daily delivery without server downtime

Accomplishments that we're proud of

  • Built a working MVP in a short time — integrated News API + LLM + daily delivery pipeline within ~1-2 weeks.
  • Created meaningful startup ideas — not generic AI output, but ideas clearly linked to current tech trends.
  • Prompt design success — tuned prompts that balance creativity with relevance, minimizing junk output.

What we learned

  • News API is a cleaner, faster alternative to manual scraping for trend data.
  • Prompt design is everything — even the best LLM will generate generic ideas unless guided well.
  • It’s important to pre-filter API results to avoid noisy or irrelevant news influencing the LLM.
  • API rate limits and costs need to be planned for, especially if scaling beyond a small MVP.
  • Automating delivery to multiple platforms (Slack, Telegram, email) requires careful API handling and error management.

What's next for IdeaPulse

  • Source diversity — add more data sources beyond News API (e.g. Reddit tech threads, Product Hunt).
  • Automation — daily ideas are sent out without manual intervention, to Slack/Telegram/email.
  • Idea archive + search — build a simple web app where users can browse past generated ideas.
  • Quality control — integrate upvotes/feedback to help the system learn what makes a great startup idea.
  • Monetization — explore a premium tier with deeper idea explanations or early trend detection.

Built With

Share this project:

Updates