The Operating System for Human Memory
We're building the infrastructure layer that makes forgetting obsolete—enabling humanity to augment cognition at scale.
$1.5M Seed Round | a16z Format
From Desktop-First → Mobile-First AI
Cloud AI APIs are now cost-effective for mobile. Privacy-first mobile architecture is no longer a compromise—it's a competitive advantage.
From Keyword Search → Memory Graphs
Humans don't remember in folders. We remember through associations, emotions, and context. The next generation of memory systems must mirror how the brain actually works.
From Passive Storage → Active Recall
The cloud became a graveyard for memories. The future is proactive—your memories resurface when contextually relevant, not when you desperately search for them.
Old Category: Cloud Storage
Dropbox, Google Drive, iCloud
$50B market, commoditized, race to zero pricing
New Category: Memory OS
Dzikra
$15B emerging market, zero competition, premium pricing
Category creation thesis: Memory is not storage. Storage is passive. Memory is active, contextual, and constantly reorganizing itself. We're building the first platform that treats memory as a living system, not a dead archive.
Cloud-Native Multi-Modal RAG
We've built the first production-ready mobile-first RAG pipeline using cloud AI + cloud vector database with <50ms query latency and <2% battery impact.
Flutter
Cross-platform
0.3s
Query Latency
<2%
Battery Impact
🔗 Knowledge Graph Architecture
Entity-relationship mapping enables semantic search beyond embeddings. We index people, places, events, and their connections.
🧠 Memory Consolidation Engine
Mimics hippocampal consolidation. Short-term memories (episodic) → Long-term memories (semantic) via nightly batch processing.
⚡ Real-Time Embedding Pipeline
CLIP + BERT fusion for multi-modal embeddings. Quantized to 8-bit for mobile efficiency without accuracy loss.
🔐 Zero-Knowledge Architecture
End-to-end encryption. Even cloud sync is E2EE. We mathematically cannot access user memories.
Moat #1: Data Network Effects
Every query improves the model. User-specific fine-tuning creates personalized memory graphs that become irreplaceable. Switching costs skyrocket after 6 months of use.
Current data: 85% retention after first successful recall. 6-month users have 10× engagement vs. new users.
Moat #2: Privacy as Architecture
Mobile-first with E2E encryption isn't a feature—it's existential. Google/Meta cannot replicate without changing their entire business model. GDPR compliance is built-in, not bolted-on.
Current data: EU users convert at 2× US rate. Privacy is a growth accelerator, not a cost center.
Moat #3: Cross-Platform Lock-In
Apple and Google are platform-locked by design. We're the only memory solution that follows you from iPhone to Pixel to Mac to Windows. Cross-platform is our wedge.
Current data: 35% of beta users have 2+ devices synced. Multi-device users have 4× higher LTV.
Beachhead: US Parents (35-45 years old)
20M
Addressable Users
$1.9B
Market Size
3.2x
Willingness to Pay
Why Parents?
Expansion Path
🎯 Year 1-2: Consumer Memory App (Now)
Photo/audio/text indexing. Natural language search. Freemium SaaS. Mobile-first.
🎯 Year 3-4: Proactive Memory Assistant
Context-aware surfacing. "Here's what you discussed with John last time." Predictive recall. Ambient AI.
🎯 Year 5-6: Memory API / Developer Platform
Dzikra SDK. Third-party apps plug into memory layer. WhatsApp → Dzikra, Slack → Dzikra. Platform play.
🎯 Year 7-10: OS-Level Memory Integration
Pre-installed on Android/iOS (partnership or acquisition). Memory becomes a core OS feature like Photos or Contacts. Exit: $5B+ acquisition by Google/Apple/Microsoft or $10B+ IPO as memory infrastructure company.
Google's Blind Spot
Business model = cloud + ads. They mine user data for targeting. We use privacy-preserving cloud AI. Privacy is existentially incompatible with their model.
"Google Photos will never be E2EE because ad targeting requires data access." – a16z memo, 2022
Apple's Limitation
Walled garden strategy = iOS-only. They will not build cross-platform. 54% of US users have mixed ecosystems (iPhone + Windows). Apple forfeits this segment by design.
Our wedge: Be the memory layer that works everywhere.
Strategic Positioning: The "Switzerland of Memory"
We're platform-agnostic, privacy-first, and open by design. Big Tech can't replicate this without regulatory risk (antitrust, GDPR). Our neutrality is our moat.
Technical Founder
Ex-Google ML Engineer. Built AI pipelines for mobile apps. Expert in Flutter, Golang, RAG architecture.
Deep expertise: RAG, vector databases, LLM quantization, mobile ML
Product Founder
Ex-Meta PM. Shipped memory features to 2B users (Memories, Stories, etc.).
Deep expertise: Consumer product, privacy design, growth
Founder-Market Fit
We've both experienced the pain of losing irreplaceable memories (lost voice messages, deleted photos). This is personal. We're solving our own problem first.
10-year friendship. Complementary skill sets (tech + product). We're in this for the long haul—this is a 10-year company, not a 2-year flip.
500
Beta Users
25
Paying ($8/mo)
85%
6mo Retention
15%
WoW Growth
Unit Economics (Actual, Not Projected)
CAC: $15 (organic referrals)
LTV: $100 (20mo avg retention)
LTV:CAC: 5:1
Gross Margin: 81%
Payback: 6 months
18-Month Milestones (Series A)
MAU: 100,000
ARR: $500k
NRR: >110% (expansion revenue)
Burn Multiple: <0.5
Series A Valuation: $50M+
$1.5M Seed
$10M post-money (15% dilution)
18-month runway to Series A
Use of Funds
60% Team (3 eng, 1 PM)
20% Infra (AI compute)
15% Marketing (growth tests)
5% Ops
Why Now?
Cloud AI pricing dropped 90% making mobile-first viable
Privacy regulation tailwinds (GDPR)
Big Tech moving slow (18mo lead)
We're building the memory layer for the next billion people.
This is infrastructure-scale. This is inevitable. And we're 18 months ahead.