All Systems
SYS_14 / 15SHIPPED • Public Dual Repo
AI SYSTEMSPublic • TypeScriptComputer Vision & Adaptive AI
Sentio (Chess Agent)
Emotion-Adaptive AI Chess Coach with Live Camera Recognition
AI chess coach combining live camera chessboard recognition, facial emotion recognition, emotion-adaptive AI difficulty, and conversational LLM move explanations across chess-agent-v2 and chess-agent.
TypeScriptComputer VisionLLMs
System ID
14
chess-agent-v2
Vision Pipelines
M-012 Streams
Board + Facial Emotion
Difficulty
M-02Adaptive
Emotion-Driven AI
Coaching
M-03LLM Chat
Move Explanations
Repos
M-04v2 + v1
TypeScript
Section 01
Problem Statement
Static chess engines neither read physical camera boards nor adapt coaching tone and playing strength to a learner's real-time frustration or confidence.
Section 02
System Constraints
- >Detect live board state and player facial emotion concurrently
- >Translate engine evaluation shifts into conversational LLM explanations
Section 03
System Architecture & Dataflow
Topology // 4 Nodes • 3 Edges
01
Camera Board & Face Input
Frames
02
Board + Emotion Recognition
Difficulty Target
03
Emotion-Adaptive Chess AI
Move Context
04
LLM Move Explanations
Section 04
Key Engineering Decisions
- 01Paired physical camera chessboard recognition with facial emotion detection to modulate AI difficulty dynamically.
- 02Added conversational LLM coaching to explain candidate moves and tactical ideas.