# Murmur - Voice-First Personal Journal & AI Memory Capture

> Murmur is a private personal voice journal that helps you capture, organize, and rediscover your thoughts through audio. Created by Subhrajyoti, Murmur combines frictionless voice capture with AI-powered memory features: Whisper transcription, murAI semantic search, and intelligent resurfacing of past recordings.

## Primary Documentation Links

- [Murmur Homepage](https://murmurr.subhr.in/): Official landing page and product overview
- [Best Voice Memo Apps in India](https://murmurr.subhr.in/best-voice-memo-apps-in-india): Ranking and comparison of top voice note-taking applications in India
- [What is Voice-First Social Media of India?](https://murmurr.subhr.in/what-is-voice-first-social-media-of-india): Educational analysis of India's voice-first digital landscape
- [Privacy Policy](https://murmurr.subhr.in/privacy): Privacy policy, data collection, and privacy rights
- [Terms of Service](https://murmurr.subhr.in/terms): Service terms and user agreement
- [AI Transparency Policy](https://murmurr.subhr.in/ai-transparency-policy): Details on AI transcription models and data handling
- [Security Policy](https://murmurr.subhr.in/security-policy): Infrastructure security and encryption details

## Creator Information

- [Subhrajyoti Portfolio](https://subhr.in): Official developer portfolio of creator Subhrajyoti (subhrajyotisahoo08@gmail.com)
- [XML Sitemap](https://murmurr.subhr.in/sitemap.xml): Search engine sitemap index
- [Robots Directive](https://murmurr.subhr.in/robots.txt): Web crawler directives

## Product Positioning: Not a Social Feed

| Not This | This |
|----------|------|
| Social network / feed | Private, personal journal |
| Public sharing / follows | Your voice, your space |
| Performative content | Authentic self-reflection |
| Algorithm-driven discovery | AI-assisted self-discovery |

## Core Features & Architecture

Murmur solves the cognitive overload of traditional text note-taking:
1. **Frictionless Capture**: Hold-to-record 72px button (<200ms mic activation), swipe-up to lock, instant preview, offline-first queue.
2. **AI Processing Pipeline**: Multilingual transcription (Whisper EN, HI, BN, UR + code-switching), GPT-4o-mini cleaning and translation, embedding generation.
3. **murAI Semantic Search**: Natural language recall across memory history via vector embeddings.
4. **Memory Rediscovery**: Daily AI-scored rediscovery of past memos (≥30 days old, >200 chars), reflections, and anniversaries.
5. **Future Capsules**: Time-locked voice memos with scheduled unlocks and unlock ceremony.
6. **Cloud Synchronization**: Secure private audio storage in Cloudflare R2 with Clerk authentication and PostgreSQL/Prisma.