Skip to content

Home

Jump to bottom
Элшан Алиев edited this page Apr 18, 2026 · 1 revision

Memory Core Enterprise v3.0 Wiki

Welcome to the official documentation of Memory Core Enterprise — the canonical implementation of the MEMORY-PRIORITY-1 protocol under the A-CODE (AC-2026-INF) standard.

🚀 Quick Navigation

Page Description
Architecture System design and component overview
API Reference Complete REST API documentation
Deployment Guide How to deploy Memory Core in 1 day
Compliance Legal compliance (152-ФЗ, ФСТЭК, ГОСТ)
Integration Examples Code examples for GigaChat, YandexGPT, DeepSeek
FAQ Frequently asked questions

📌 What is Memory Core?

Memory Core is an on-premise microservice that gives LLM assistants long-term memory across sessions.

  • 50x token cost reduction — compressed cognitive profile
  • 85-90% personalization accuracy — semantic vector search
  • < 1 day deployment — Docker, single command
  • Full legal compliance — 152-ФЗ, ФСТЭК, ГОСТ

🏗️ Architecture Overview

User → LLM (GigaChat/YandexGPT/DeepSeek) → Memory Core API ↓ PostgreSQL + pgvector ↓ Ollama (Llama3) ↓ Cognitive Profile (24h refresh)

📡 API at a Glance

Endpoint Method Description
/remember POST Store a fact
/recall POST Semantic search
/summary GET Get cognitive profile
/summary/refresh POST Force profile update
/health GET Health check

🛡️ Compliance

Memory Core is fully compliant with:

  • ФЗ-152 (Personal Data Localization)
  • ФЗ-149, ФЗ-187 (Information Security)
  • Приказ ФСТЭК №21 (Access Protection)
  • ГОСТ Р 57580.1-2017 (Banking Security)

📋 View Full Compliance Statement

👨‍💻 Author

Elshan Aliev — Architect of A-CODE (AC-2026-INF) and creator of Memory Core.

© 2026 Memory Core Enterprise. All rights reserved.

Clone this wiki locally