Flashback AI [BETA]

About

Flashback is a Privacy-First AI memory app that listens to your life stories as you share them in audio chats with our voice assistant, and Flashback turns them into structured Memory Orbs you can own, verify, and control without surrendering your personal data to ad-driven platforms. You talk; Flashback organizes memories around people, time, and emotion, and preserves them as a private archive designed to train your future AI twins. ⸻ How it works • Tap the mic & talk Share a moment from today, a childhood story, a call with your mom, or something you don’t want to lose. No typing. Your data stays private and tied to your identity, not a public profile. • AI extracts the “Flashbacks.” Flashback detects the moments that matter, who they were about, what happened, when/where, and why it mattered then saves them as structured memories in your private timeline. • Memories become Memory Orbs Each moment becomes a Memory Orb: an emotion-rich capsule containing story, metadata, and relationships. Orbs are private by default and designed to feel like a safe vault, not a feed. ⸻ What are Memory Orbs? Memory Orbs are living containers for your life: • A story + context of a real moment • Linked to the people in your life • Organized by time, emotion, and relationships • Stored in one private place you control anytime access No public timeline. No engagement farming. No ad incentives. ⸻ Why Web3 users care • Ownership, not platforms Your memories aren’t “content” for someone else’s business model. Flashback is built around a user-owned data ethos. • Transparency & consent You can see what’s being processed, what’s stored, and what stays private. No silent data extraction. • A foundation for future AI twins Every Memory Orb becomes part of a growing personal dataset that can later power your AI twin trained from your own memories and relationships. • Monetization on your terms (optional) If you ever choose to share or license outputs, it’s designed to be opt-in, auditable, and aligned with creator ownership, not “free training data.” ⸻ Privacy, by default • Private by default: no discover feed, no public timeline • Control: edit/delete or export your orbs and recordings • No ad profiling: your memories aren’t used to build ad graphs • Secure processing: built to treat memories as sensitive data ⸻ Who it’s for • Web3 natives who want user-owned personal data • Parents and adult children preserving family stories • Anyone who prefers speaking over journaling • People building a private archive that can later train an AI twin

What’s new in 0.4.10

1. UI Design Improvements 2. Memory module implementation for MeeMaw 3. Fine tunes applied for local model inferencing 4. Bug fixes