Short Summary: Use Reflex Speaking AI for its intended learning purpose, protect your account, carefully review content before sharing, confirm transactions before making payments, and contact support if you encounter any issues.
1. Scope of Application
This policy applies to the entire experience on the phanxa.com website (Reflex Speaking AI), including login, English speaking practice, Reflex Drill mode, viewing results, practice history, personal profile, Premium plan management, question bank, and lessons.
When you access, register, log in via Google, or continue using the service, you confirm that you have read, understood, and agree to be bound by the terms in this document.
2. User Accounts
You are responsible for the security of your account, device, browser, and login session. Do not share account access with others if it creates risks for data, payments, or learning progress.
Profile information such as name, email, Premium status, learning progress, activity statistics, and class data is used to operate the service and personalize the learning experience.
Reflex Speaking AI reserves the right to temporarily lock or terminate access if an account shows signs of fraud, spam, system sabotage, unauthorized use of automated tools, or violation of law.
3. Google Login
Google is the only supported login method. When logging in via Google, we only receive the information necessary to identify users such as email, display name, and identification code.
We do NOT store your Google password. Login sessions are managed via secure tokens in the browser.
After logging in, the system may automatically redirect you to the page you were previously working on, such as a shared lesson link or the Premium upgrade page.
You can log out at any time. Tokens and local session data will be removed from the browser upon logout.
4. English Speaking Practice
You record your answers using the browser microphone for English questions. Audio is processed (edge trimming, silence removal, volume normalization, conversion to 16kHz WAV format) and then sent to AI for evaluation.
Audio is temporarily stored in IndexedDB on your browser and is NOT permanently stored on our servers.
You are responsible for the content of the audio you record. Do not record harmful, illegal, privacy-infringing, or intellectual property-violating content.
Each question has a 15-second time limit to answer. You can skip questions if you wish.
5. Reflex Drill Mode
Reflex Drill is a fast-paced practice mode with limited reaction times: Easy (10 seconds), Medium (7 seconds), Hard (5 seconds).
Each drill consists of 3 questions. The recording and AI evaluation process is similar to regular practice.
Drill results are saved separately in your practice history.
6. Question Bank and Lessons
Questions are organized by topic and CEFR level (A1–C2). There are two types of questions: "answer" (responding to a question) and "ask" (asking an appropriate question for the situation).
Each question includes: English content, Vietnamese translation, sample answers (EN + VI), and expected keywords.
Lessons group questions into structured practice sets. Questions may be created by AI or manually by administrators.
You can report issues with questions if you find incorrect or inappropriate content.
7. AI Evaluation and TTS
AI evaluates skills: pronunciation, vocabulary, grammar, fluency, and confidence (each criterion scored 0–100).
AI also provides: audio recordings, comments, translations, and suggested answers.
The TTS system (text-to-speech) plays question audio using Browser SpeechSynthesis, ElevenLabs, or MiMo depending on configuration.
AI results are for reference only — they do not guarantee absolute accuracy. You should not rely solely on AI evaluations for official exams.
AI may occasionally produce errors, biases, or inappropriate content. Please verify and take responsibility when using AI-generated content.
8. Reward System and Levels
Each practice answer costs 1 point (Free plan: 100 points/day, Pro plan: unlimited).
XP (experience points) earned: 30 XP per practice question, 25 XP per drill question.
Streak tracking: consecutive days of practice. There are streak protection tokens to prevent losing your streak when taking a day off.
CEFR level progression: A1 → A2 → B1 → B2 → C1 → C2 (based on cumulative XP).
19 achievements unlocked automatically based on milestone accomplishments.
9. Payments and Premium Plans
Current Premium plans:
• Free Plan: Free, 100 points/day
• Pro 3 months: 189,000 VND, unlimited
• Pro 6 months: 349,000 VND, unlimited
• Pro 1 year: 629,000 VND, unlimited
There is no integrated payment gateway — Premium plans are managed manually by administrators. Contact admin to upgrade and make payments.
Prices, benefits, and terms may change over time.
Expired Pro plans will automatically revert to the Free plan.
10. Refunds and Issue Resolution
Refund requests are reviewed on a case-by-case basis, depending on activation status, usage duration, actual technical issues, and payment receipts.
We may refuse refunds if the plan has been substantially used, the transaction shows signs of fraud, the user has violated terms, or the request lacks sufficient verification information.
Contact admin with the following information: order code, transfer time, amount, and receipt photo for reconciliation.
We are not responsible for fees separately charged by banks or intermediaries.
11. Data We Collect
Account data: email, display name, user ID, role (user/admin), registration status.
Practice data: questions, answers, audio recordings, evaluation results, scores, skill analysis, session metadata.
Gamification data: XP, streak, level, achievements, daily points.
Technical data: login token, browser information, request timestamps, system errors.
LLM logs: token usage, request/response content (truncated), audio metadata (size), latency — automatically deleted after 14 days.
We do NOT store: Google passwords, full bank card information, permanent audio recordings.
12. Purpose of Data Usage
Provide, operate, and personalize the English speaking practice experience.
Sync learning progress across devices.
AI evaluation of your spoken answers.
Generate question audio via TTS.
Authentication, security, spam prevention, rate limiting.
Premium plan management and admin operations.
LLM performance monitoring (14-day logs).
Improve the product based on aggregated usage metrics.
We do NOT use your data for cross-application targeted advertising.
13. Third-Party Data Sharing
We only share data when necessary:
• Google (OAuth): email, name — for login only
• OpenCode AI: audio (base64), text, conversation history — for AI evaluation, translation, question generation
• ElevenLabs (optional): text — for TTS audio generation
• MiMo/Xiaomi (optional): text — for TTS audio generation
We do NOT sell personal data to third parties.
Data is only shared to the extent necessary to operate the service.
Data may be provided in response to valid legal requests from authorities.
14. Storage and Security
SQLite database on the server with access controls, authentication tokens, and rate limiting.
Audio stored in IndexedDB on the browser (client-side), NOT on the server.
LLM logs: 14-day TTL with a 5,000-line limit, automatically deleted.
Daily points: auto-cleaned after 7 days.
No online system is completely secure — protect your devices and do not share login credentials.
Data is retained for as long as necessary to provide the service, resolve disputes, and comply with legal obligations.
15. Cookies and Browser Data
Cookies/localStorage are used for: login sessions, UI preferences, temporary practice data, return paths after login.
No tracking cookies, no advertising cookies, no third-party analytics.
Disabling cookies may prevent login, session sync, and some features from working correctly.
16. User Rights
View and edit profile information (display name) directly in the application.
View practice history, scores, skill analysis, and achievements.
Contact admin to request viewing, editing, or deleting personal data.
Transaction records are retained for 10 years in accordance with Vietnamese accounting document storage regulations.
There is currently no self-service account deletion feature — please contact admin.
17. Prohibited Activities
Attacking, scanning, sabotaging, bypassing limits, exploiting vulnerabilities, interfering with the API, or using unauthorized bots.
Bulk copying unauthorized data, reselling accounts, reselling Premium content, or sharing paid access.
Posting offensive, threatening, fraudulent, copyright-infringing content, or content that disrupts the learning environment.
Impersonating identities, creating bulk accounts, manipulating leaderboards, gaming the reward system, or exploiting system bugs.
Using AI-generated content for scams, academic fraud, or harm.
18. Intellectual Property
The Reflex Speaking AI name, interface, source code, design, features, and curated content are the property of Reflex Speaking AI or its licensors.
You retain rights to your own created content (recorded audio), but grant us the right to store, process, display, and technically reproduce it to the extent necessary to provide the service.
AI-generated content (questions, translations, comments) is the property of the platform.
19. Service Changes and Interruptions
We may update, temporarily suspend, limit, or discontinue certain features for maintenance, improvement, security, or regulatory compliance.
In the event of a failure, we will make efforts to restore the service and data, but we cannot guarantee that all temporary data or unsynchronized states can always be recovered.
20. Limitation of Liability
Reflex Speaking AI is a learning support tool and does not guarantee exam results, grades, certificates, or specific achievements.
We are not liable for damages arising from user input errors, accidental content sharing, use of insecure devices, use of unverified AI content, or violation of terms.
AI evaluations are for reference only and may contain errors.
21. Minors
The service is intended for users aged 13 and above due to AI and audio recording features.
Users under 18 should have parental or teacher supervision.
AI and audio recording features require appropriate consent.
Parents/teachers are responsible for supervising the use of minor users.
22. Policy Updates
We may update this policy when the product, operational requirements, or legal regulations change. New versions take effect when published on the website.
If significant changes materially affect user rights, we will display a notification in the application.