Hey HN,
I’m Suraj, founder of Inspirt Labs. After losing someone close, I wanted more than photos and voice notes—I wanted real conversation. So I built a service that turns the fragments we already have (pics, voice clips, text threads, quirks, life stories) into a living, conversational replica you can speak with anytime.
Personalized AF – every model is fine-tuned on the exact memories you provide, so it sounds and responds like your person, not a generic chatbot.
Multimodal – voice, text, and an optional photoreal avatar.
Ethically grounded – we require consent from legal next-of-kin, keep replicas private by default, and allow permanent deletion with one click.
Built in a week – OpenAI + a thin orchestration layer + ElevenLabs for voice. MVP is live today.
Demo: https://inspirtlabs.com (upload a short sample set and chat in under 2 min) Tech notes: repo + infra costs + fine-tuning pipeline in the README. Looking for feedback: latency, uncanny-valley moments, and pricing model ideas. Rip it apart!
Thanks for reading—feel free to ask me anything.