Table of Contents
Welcome to Violet Avatar
Violet Avatar is an interactive platform powered by AI Avatars with advanced Flow + Intent architecture. Customizable, emotionally intelligent, and scenario-aware digital humans that adapt to your needs with precision and empathy — from education to enterprise.
Five Application Scenarios
Education & Training — Language learning partners and automated oral proficiency testing. Create immersive learning environments.
AI Interviewer — Automated candidate screening and mock interview practice. Standardize your hiring process.
Marketing & Sales — 24/7 Brand Ambassadors and e-commerce guides. Engage customers with personalized interactions.
Companionship — Elderly care assistants and emotional support avatars. Provide presence and empathy.
Customer Service — Intelligent support agents with sentiment analysis. Detect abnormal emotions and escalate when needed.
Cloud Collaboration & AI Secretary — Share files, chat, and make multi-party calls with a single link. AI-powered real-time translation and meeting summaries.
AI Tour Guide — Point your camera at an exhibit and let AI tell its story. Smart guided tours with real-time visual recognition and voice narration.
Quick Start
Login & Register
Click the 'Login' button at the top right. Supports quick login via Google or Email registration. Once logged in, your learning progress, created lessons, and characters will be synced across all devices.

Language Switching
The platform supports Traditional Chinese and English interfaces. Click the language icon (🌐) at the top right to switch. The system will automatically remember your language preference.
Learning Center
The Learning Center is the platform's core feature, bringing together community-created interactive lessons. You can search, browse, and start immersive learning sessions with AI teachers.

How to Create a Lesson?
Utilizing advanced AI technology to automatically convert YouTube videos into structured interactive lesson plans. Suitable for language learning, knowledge sharing, and more.
Input Video
Paste a YouTube video URL, and the system will automatically download subtitles and analyze segments to extract key learning points (approx. 2-5 mins).
Edit Content
Enter edit mode to fine-tune video segment start/end times, modify AI-generated summaries, and choose the most suitable AI teacher from the character library.
Publish
Set the lesson visibility. Public lessons will be shared with the community, while private lessons are for your own use or your class students only.
How to Study a Lesson?
Click any lesson card to enter the learning page. The AI teacher will guide you step-by-step through key vocabulary, sentence patterns, and cultural knowledge based on video segments, with real-time voice conversation practice.
Teaching Center
Classroom management tools designed for teachers and educational institutions.
- Class Management
After creating a class, you will get a dedicated teacher dashboard to view all student learning hours, lesson completion rates, and interaction performance.
- Invite Students
Share the 6-digit invite code with students. When students join, the system prompts for data consent to ensure privacy compliance. Teachers can manage the student list at any time.
- Learning Analytics
Teachers can view detailed learning data for each student, including practice counts per lesson, conversation duration, knowledge point mastery, and more — effectively tracking teaching outcomes.
Character Studio

Create Your AI Avatar
Supports uploading standard VRM or GLB format 3D models. You can configure unique personas, voice parameters, and emotional response modes for your character.
Voice Config
- Google TTS:Google Cloud TTS. Pros: Fast speed, high free tier usage, suitable for general conversation and long text reading.
- Azure TTS:Microsoft Azure TTS. Supports rich emotional style adjustments (e.g., excited, whispering, serious) for more natural voice performance.
- ElevenLabs:ElevenLabs AI Voice. Industry-leading realism, supports voice cloning, but with higher cost and slightly slower generation speed.
- Cartesia:Cartesia Sonic. Ultra-low latency voice model designed for real-time interaction, providing the smoothest conversation experience.
AI Interviewer
Whether you're a job seeker looking to practice or a company wanting to automate recruitment — the AI Interviewer has you covered.
Interview Practice
Choose different interviewer personas (e.g., strict HR, technical lead) and simulate real workplace scenarios. AI asks follow-up questions in real time and provides detailed feedback with scoring after the session.
Become an Interviewer
Create interview campaigns with job descriptions, questions, and scoring criteria. Share invite codes with candidates — AI conducts standardized interviews, records sessions, and generates evaluation reports for each candidate.
Evaluation Reports
After each interview, structured evaluation reports are auto-generated with per-question scores, overall performance analysis, and video playback. Companies can review all candidates' results in one dashboard.
Create or Choose
Companies create interview services with custom questions; candidates browse public interviews or enter an invite code.
Start the Interview
The AI interviewer guides the session with voice, asks follow-up questions, and records the conversation. Video recording is optional.
Get Results
After the interview, candidates receive improvement suggestions; companies get structured reports and candidate rankings.
AI Tour Guide
Point your camera at an exhibit and let AI tell its story. Combining real-time visual recognition with voice narration for an immersive smart guided tour experience.
Real-time Camera Recognition
Point your rear camera at exhibits, paintings, or architecture — AI automatically identifies the content and provides professional commentary. No QR codes or manual input needed.
Voice-guided Narration
The AI tour guide narrates the history, artistic features, and fascinating stories behind each exhibit in natural speech. Supports multiple languages for international visitors.
Custom Tour Routes
Administrators can create dedicated tour services with exhibit databases, narration styles, and guide characters. Visitors simply scan a code or enter an invite code to start.
Choose a Tour
Visit the Tour Center and select a museum, exhibition, or landmark tour service.
Point at an Exhibit
Allow camera access and point your rear camera at the exhibit. AI recognizes it instantly and starts narrating.
Listen to the Story
The AI guide narrates the exhibit's story through voice while displaying text descriptions and extended knowledge.
Cloud Collaboration
A lightweight team workspace — share files, chat, and make calls, all from a single link. No login or app installation required for guests.
Cloud Disk
Create a shared space and get a unique link. Anyone can upload, download, and manage files — perfect for team collaboration and file exchange.
Messaging
Send text and image messages directly within the share page. Push notifications ensure you never miss an important message.
Multi-party Calls
Initiate audio or video calls with push notification alerts. Supports accept, decline, and ringtone — just like a real phone call experience.
Call Controls
Toggle microphone mute, camera on/off, and switch between front and rear cameras during calls. Multi-party support with automatic call history.
How It Works
Create a Share
Create a cloud share space, or start a call directly from an existing one.
Share the Link
Send the link to anyone — guests can join without signing up or logging in.
Start Collaborating
Share files, exchange messages, and make voice/video calls — all in one place.
AI Meeting Secretary
Invite an AI secretary into any call for real-time transcription, multilingual translation, and post-call summaries. Break down language barriers in cross-national meetings.
AI Voice Bridge
The AI secretary listens to all participants and translates speech into each person's language in real time. Each language gets its own audio track. Supports Chinese, English, Japanese, and Korean. Languages are auto-detected from your settings — just speak naturally.
- Flexible Configuration
Customize features when inviting the secretary: real-time transcription, text translation subtitles, voice translation, and meeting summary. Choose which languages to support.
- Live Subtitles
Speaker names, original text, and translations appear at the bottom of the call screen in real time. Subtitles fade automatically without obstructing the video.
- Post-call Summary
After the call ends, AI automatically generates a meeting summary with key decisions, action items, and assignees. Full transcripts are available for review anytime.
Technology
At the core of Violet Avatar is a real-time voice interaction engine — from the moment you speak to the AI avatar's response, everything happens within seconds. Here are the key technologies that make it all work.
Real-time Voice Interaction Pipeline
Speech Recognition
Every word you say is recognized in real time, supporting multiple languages and domain-specific vocabulary.
Dialogue Engine
AI understands your intent and determines the best response based on conversation progress.
Voice Synthesis
AI responses are converted to natural speech instantly, with multiple voice styles and languages available.
Avatar Rendering
3D avatar lip-syncs in real time with natural blinking and head movements.
Precise Dialogue Control
Unlike typical chatbots that generate free-form responses, our dialogue engine uses structured flows to precisely control interactions (e.g., lesson steps, interview questions), while retaining natural language understanding. The result: AI that is both precise and natural.
Cross-session Memory
The AI remembers you. Three memory layers: long-term facts (your name, preferences), recent summaries (what you learned last time), and persona profile (your level and style). Persisted across sessions and automatically loaded on reconnect.
Real-time 3D Avatar
Upload your 3D model, and the AI avatar will drive lip-sync animations in real time based on speech. Natural blinking and subtle head motion play during pauses, keeping the character lifelike.
Multiple Voice Styles
Multiple voice styles to choose from: fast and stable, emotionally expressive, ultra-realistic (with voice cloning support), and ultra-low latency. Each character can use a different style to find the perfect voice.
Safety Guard Rails
Built-in content safety framework. Each character can have its own conversation rules to keep interactions appropriate. Input validation and output filtering provide dual-layer protection against inappropriate content.
Real-time Low-latency Transport
Sub-second voice latency, works in any browser without installing plugins. Real-time delivery of subtitles, translations, and status updates — zero-wait communication.
Other Applications
Interview Practice
Simulate real workplace scenarios. You can choose different interviewer roles (e.g., strict HR, technical lead), and AI will ask follow-up questions based on your answers and provide improvement suggestions.

Companion Chat
Provides 24/7 emotional support and companionship. AI characters have long-term memory to remember your preferences and past conversations, building a deeper connection.

Marketing & Sales
24/7 Brand Ambassadors and e-commerce guides. AI avatars engage customers with personalized interactions, providing product introductions and event recommendations.
Customer Service
Demonstrates enterprise-level customer service applications. Combined with RAG (Retrieval-Augmented Generation) technology, AI can accurately answer professional questions about product specs and after-sales service.