Introduction

Welcome to MindFry

Experimental

MindFry is currently in active R&D. The API is volatile. It simulates biological memory processes which may result in data inhibition (data loss from the user’s perspective) based on the system’s “mood”. Do not use for banking.

What is MindFry?

MindFry is the world’s first Ephemeral Graph Database — a substrate for machine memory that treats data as living neurons, not static records.

Unlike traditional databases that store data deterministically (0 or 1), MindFry uses a Tri-Cortex Engine based on Balanced Ternary Logic (-1, 0, +1) to simulate organic cognitive processes.

Key Differentiators

Feature	Traditional DB	MindFry
Storage Model	Permanent / TTL	Organic Decay
Relationships	Foreign Keys	Synaptic Bonds
Query Side Effects	None	Observer Effect
Data Accessibility	Always On	Mood-Modulated
Logic System	Binary	Balanced Ternary

Consciousness States

MindFry memories exist in one of three consciousness states:

• LUCID — Active, high-energy memories readily accessible
• DREAMING — Latent memories that may surface under specific conditions
• DORMANT — Deep subconscious memories, accessible only via Executive Override

The Stack

MindFry is built on a Zero-GC, Data-Oriented Design architecture:

• Engine: Rust (bare-metal performance)
• Protocol: MFBP (MindFry Binary Protocol) over TCP
• Persistence: sled (Rust-native embedded database)
• SDK: TypeScript (npm), with Python and Go planned
• License: Apache-2.0

Quick Links

• Quickstart — Run MindFry in 5 minutes
• Concepts — Core ideas explained
• Mental Model — Why MindFry exists
• Architecture — Deep dive into internals