China First

AI Education — Otowyue Learn AI

China's First FSRS × Digital Human · Fully Offline · AI Teacher with Soul

Redefining Intelligent Education

Existing AI education products share three core flaws: generic AI lacks teaching logic; institutions offer one-size-fits-all content; traditional classrooms have 40:1 student-teacher ratios with no personalization.

Otowyue Learn AI puts emotion sensing at the core, FSRS scientific memory as the engine, and VRM digital humans as the interface — creating an “AI teacher with soul” that makes every student feel understood and seen.

❤️ Human Warmth Emotional companion teaching
Emotion-Aware Adaptive learning strategy
Personalized Individual knowledge graph
🔒 Fully Offline Data never leaves device

Five Core Innovations

🎭

Emotion Sensing Engine

Text / voice / expression three-channel multimodal fusion. 92% emotion recognition accuracy, <800ms response, 8 emotion types, dynamically adjusting teaching pace and methods.

92% Recognition Accuracy
China First
🧠

FSRS × Digital Human Integration

Based on MIT open-source FSRS scientific memory algorithm, precisely calculating optimal review timing. China's first educational AI product to deeply integrate FSRS with VRM digital humans.

93%+ Retention Rate · 3x Learning Efficiency
🗺️

Knowledge Graph Navigation

Visualized knowledge pathways covering all K-12 subjects. Prerequisite/dependency mapping, automatic bottleneck identification, optimal learning path planning.

Real-time Mastery Tracking
👥

Four-Role AI Collaboration

AI Teacher (instruction) / AI Assistant (Q&A) / AI Advisor (learning analytics) / AI Counselor (emotional support) — four roles collaborating to simulate a real teaching team.

Multi-dimensional Coaching
🤖

VRM Digital Human Interaction

Humanized VRM digital teacher, natural language dialogue, MFA phoneme-level pronunciation evaluation, real-time emotion expression mapping, <2s response, 11+ LLM compatible.

11+ LLMs · <800ms End-to-End Voice

Key Metrics

92% Emotion Recognition Accuracy
93%+ FSRS Memory Retention
3x Learning Efficiency Gain
<800ms End-to-End Voice Response
11+ Compatible LLM Models
60,000+ Lines of Production Code

Fully Offline Architecture Advantages

Tech stack: Ollama + Sherpa-ONNX + MeloTTS + VRM + FSRS

🔒

Data Never Leaves Device

Video stays on device, only emotion labels are transmitted. Meets school intranet security requirements, compliant with GB/T standards.

📡

Works Without Internet

Fully offline deployment. Rural and remote schools can use it too, advancing educational equity. Covering 200+ rural schools.

🔓

Open Source & Customizable

Apache 2.0 license. Free to customize without lock-in. School IT teams can independently develop extensions.

🔄

LLM-Agnostic

Supports 11+ large models with no vendor lock-in. Seamlessly switch to better models as they become available.


Use Cases

🏫
K-12 Schools
ToB institutional edition, ¥50K–200K/year, customized knowledge graphs, batch student management, analytics dashboards
Apply for Pilot →
👨‍👩‍👧
Home & Personal
ToC family edition, ¥39–99/month, personalized learning plans, emotional companion mode, parent monitoring panel
Learn More →
🌾
Education Bureaus / Rural
ToG government edition, ¥500K–2M/project, fully private deployment, regional data dashboards, education equity programs
Contact Us →

Contact Us →

Every Child Deserves an AI Teacher with Soul

Apply for school pilot, family trial or government partnership

Contact Us →