Skip to content

MemCtrl v1.0.0 — Cognitive Memory Runtime for AI Agents

Choose a tag to compare

View all tags
@KJ-AIML KJ-AIML released this 23 May 14:17
v1.0.0
0d64f03

MemCtrl v1.0.0 — Cognitive Memory Runtime for AI Agents

Overview

MemCtrl is an operating system for long-lived agent memory — hierarchical, explainable, and self-managing. It replaces passive vector dumps with human-like memory layers that remember, forget, consolidate, and explain their reasoning.

What's Included

Core Architecture

  • Hierarchical memory layers — Project (forever), Session (7 days), User (90 days)
  • Tree-based retrieval — LLM reasons over memory structure, not vector similarity
  • Reasoning traces — Every answer shows its exact path: root -> project -> auth -> jwt
  • Automatic consolidation — Session memories merge into project knowledge via trigger rules
  • Security-first — Secrets, API keys, and PII are redacted before storage

Integrations

  • LangGraph — MemCtrlSaver checkpoint saver + MemoryNode for agent workflows
  • MCP — Stdio transport server for IDE integration
  • Claude Code / Cursor / Kimi Code / Codex — SKILL.md registration via memctrl install

CLI Commands

  • memctrl init — Create .memoryrc
  • memctrl add — Store memory with layer, tags, confidence
  • memctrl query — Retrieve with reasoning trace
  • memctrl heatmap — Visualize memory distribution
  • memctrl timeline — Chronological memory events
  • memctrl tree — Hierarchical tree view
  • memctrl trigger-cmd — Fire automation rules
  • memctrl audit — Complete trigger log

Demos & Benchmarks

  • examples/coding_agent_demo.py — Multi-session agent simulation
  • examples/killer_demo.py — Bug prevention across sprints (the "holy sh*t" moment)
  • examples/langgraph_integration.py — LangGraph usage patterns
  • benchmarks/retention_benchmark.py — Measurable retention vs naive baseline

Visualizer

Installation

pip install memctrl

Quick Start

memctrl init
memctrl add "we use FastAPI + PostgreSQL" --layer project
memctrl query "what is our stack?"
# Trace: root -> project -> tech_stack -> FastAPI + PostgreSQL

Benchmarks

Metric Vector RAG MemCtrl
Context Retention (10-turn) 62% 91%
Retrieval Explainability 0% 100%
Memory Management Overhead Manual Zero ops
Long-Horizon Task Success 45% 78%

Links

Thanks

To everyone who shaped MemCtrl's direction. This is just the beginning.

Full Changelog: https://github.com/KJ-AIML/memctrl/commits/v1.0.0

Assets 2
Loading