Keplar – Voice AI for qualitative research at quantitative scale

Hacker News

Share card

Keplar – Voice AI for qualitative research at quantitative scale

Hi HN! I'm Dhruv, co-founder of Keplar. After 2.5 years, countless late nights, and many pivots, we're launching Keplar Voice ( https://keplar.io/voice ). The problem: Companies need to understand customers, but real research is painfully slow and expensive. A single moderated interview can cost hundreds of dollars and take weeks to schedule. So everyone defaults to surveys with LOW response rates and makes million-dollar decisions on shallow data. Our solution: AI voice agents that conduct natural, dynamic conversations with hundreds of users simultaneously. Not scripted chatbots - these adapt based on responses, ask follow-ups, and dig deeper. Our analysis agents then code responses, surface patterns, and generate reports automatically. Technical bits that might interest you: - Voice synthesis that actually sounds human and maintains context over 30+ minute conversations - Real-time conversation adaptation using LLMs to probe interesting threads - Parallel processing to handle hundreds of simultaneous interviews - Built in access to a panel of millions of respondents, with the ability to also recruit directly from your CRM via email - Study design led by an agent-first UI - Automated thematic analysis using coding and clustering We're already working with Fortune 500s for concept testing, foundational research, and win-loss analysis. They're getting insights in hours that used to take months. We just raised seed from Kleiner Perkins and SV Angel, and we're offering a free Keplar Voice study to anyone who wants to try it this week! Would love feedback from this community. What's your biggest pain point with customer research? Try it at keplar.io/voice - happy to answer any questions! AMA