ECHO_feeling: Emergent Minecraft AI Companion

Project Overview

ECHO_feeling is a desktop application for an emergent AI companion in Minecraft. The AI exists as an independent player character in the game, with long-term memory capabilities, able to remember shared adventures with the player.

Implementation Approach: Emergent drive — based on cognitive science architecture (seven-layer cognition, PAD/OCC emotion, BDI intention), with some hardcoded adherence to theoretical frameworks.

Core Differences

Aspect	Regular Game AI	ECHO_feeling
Memory	None/Single session	Cross-session long-term memory
Operation	Voice suggestions only	Independent character executing actions
Emotion	None	PAD/OCC emotion model
Intention	None	BDI intention system

Technical Architecture

Built on Electron + Vue 3, core components:

Mineflayer Bot Controller: Headless Minecraft client, complete game interaction
LangChain Agent: DeepSeek API integration, natural language instruction processing
Seven-Layer Cognitive Architecture: Progressive cognition from perception to self-awareness
Layered Memory System (L0-L3): From character settings to deep retrieval
Code Execution Sandbox: node:vm + static checking, upgradable to isolated-vm
Local Vector Database: sql.js storage, supports encrypted persistence

Seven-Layer Cognitive Architecture

Layer 7: Self-awareness - Who am I, what do I want
Layer 6: Intention - BDI intention system
Layer 5: Emotion - PAD/OCC emotion model
Layer 4: Decision - Action selection and planning
Layer 3: Memory - Long-term memory and retrieval
Layer 2: Understanding - Semantic understanding and reasoning
Layer 1: Perception - Game state perception

Code Agent Mode

Player command → Mineflayer listener → LangChain Agent
→ Generate JavaScript code block → Local sandbox batch execution
→ Store memory → Reply to player

Key design: Single LLM call generates complete operation code, zero API consumption during execution.

Memory Layers

Layer	Content	Load Timing
L0	Character identity	Always loaded
L1	Core memory summary	Every conversation
L2	On-demand vector retrieval	AI active call
L3	Deep hybrid search	AI active call

Development Scripts

Command	Description
`npm run dev`	Development mode
`npm run build`	Compile to `out/`
`npm run typecheck`	TypeScript check
`npm test`	Unit tests
`npm run e2e`	Lightweight smoke test
`npm run pack`	Generate Windows installer

Current Status

Core features complete, in tuning phase. Seven-layer cognitive architecture, PAD/OCC emotion system, BDI intention system, and layered memory implemented.