About

A founder-led product for families who know the stories matter.

Alfaaz turns elders' WhatsApp voice notes into a family archive, currently in Hindi, Indian English, and Marathi, with Hindi/Hinglish code-mixing supported. The original audio stays with the story. Built in 2026 by Pulkit Mendiratta and Shivli Gupta.

What Alfaaz is

Alfaaz is a WhatsApp-based memoir service for Indian families. A family member signs up, adds their elder's WhatsApp number, and starts the first conversation when they are ready. Alfaaz then asks gentle voice-note questions, one at a time, in the elder's own language. Every reply is transcribed and organized into a growing family archive with the original audio kept alongside it.

It is built specifically for Indian elders. It runs inside WhatsApp, so there is no new app to learn. It works in Hindi, Indian English, and Marathi today, with natural Hindi/Hinglish mixing supported, and it moves at the elder's pace instead of a schedule. Telugu is in review. The archive is open to invited family members and is being designed toward export and print as it matures.

Why Alfaaz exists

Most people I know carry the same quiet regret. They meant to sit down with their grandparents. They meant to ask about Partition, the first job, the village before everything changed. They kept meaning to. Then they ran out of time.

I grew up in a joint family where the stories were everywhere. Dinner-table stories, festival stories, stories told half in Hindi and half in English, full of people I'd never met and places that no longer existed. My grandparents were the living archive of all of it. Like most families, we assumed those stories were somehow permanent. That we'd always have time. We didn't.

The problem isn't that families don't care. It's that the conditions to record a life story almost never line up by themselves. You need two people free at the same time, across a time-zone gap, in the right mood, with a recorder running, asking the right questions. That never happens. Work, health, logistics, the ordinary weight of daily calls. Something always gets in the way.

Alfaaz starts from a different premise. Indian elders already use WhatsApp voice notes. They send them to grandchildren, to siblings, to friends from thirty years ago. The habit is there. The interface is trusted. The only missing piece is someone asking the right question at the right moment, often enough that a lifetime of stories actually gets told.

That is what Alfaaz does. Not as a replacement for the calls you should still make, but as something that works when the calls don't happen. An AI listener that asks one thoughtful question, waits however long it takes, and turns the answers into something your family can keep.

WhatsApp-native

No new app. No extra login. No account for the elder. Alfaaz is useful only if the first conversation feels easy. For Indian elders over 60, WhatsApp voice notes already feel easy.

India-focused

Hindi, Indian English, and Marathi are first-class product concerns. Hindi/Hinglish mixing is handled as part of the Hindi experience. Telugu is in review, and more Indic languages follow in the order our waitlist asks for them.

Voice-preserving

The archive is not just text. The original audio stays with each story, because a family voice carries emotion, cadence, and language that a transcript alone cannot hold. What you keep is the voice itself.

Product principles

  • The elder's experience is the product. Every design call is weighed from the elder's side first. If it adds friction for them so things are easier for us, we pick their ease.
  • Family stories are not training data. We will not use your family's recordings, voices, or stories to train AI models. The stories belong to your family.
  • Honesty over polish. We say plainly where the current infrastructure is rough, even when it isn't flattering. Vague trust language is not trust.
  • The archive must outlast the platform. We are building toward export and portability. Your family should never be locked into Alfaaz to reach what your elders shared.

What we will not do

  • Force elders into a new workflow just because it is easier for the product team.
  • Use family stories as training data for AI models.
  • Hide behind vague claims about privacy or data handling.
  • Expand into thin, search-first content that says little and helps nobody.
  • Lock families into Alfaaz to access recordings they shared with us.