Answers to the most common questions about using VOXS — including our security architecture, data handling practices, and compliance posture.
No. Audio is never recorded, transmitted, or stored. Speech recognition runs entirely within the trainer's browser using the device's built-in Web Speech API. The audio does not leave the trainer's device at any point. Only the transcribed text — after it has been converted to text locally — is sent over the network for translation.
All connections — between trainers, the platform, and students — use TLS 1.3, the current industry standard for transport layer security. There are no unencrypted communication paths in the system. Data at rest is handled by our infrastructure provider's encryption standards.
None. Students are not required to provide any personal information. There is no registration, no account, no name, no email address, and no device identifier collected or stored. Students scan a QR code, select a language, and join. They leave no data trail whatsoever.
Session data is ephemeral. Translated text is broadcast in real time and is not persisted after the session ends. There is no transcript archive, no session log, and no retained content. When the session is closed, the data does not exist anywhere in the system.
Yes. Because no personal data is collected from students and session content is not retained, the platform has a minimal data processing footprint. Trainer account data is handled in accordance with GDPR requirements. There is no personal data to subject access request, rectify, or erase for student participants — because none is collected in the first place.
Each student receives a cryptographically scoped authentication token when they join. This token grants access to one session channel only — the specific session they joined. It cannot be used to access other sessions, the trainer interface, or any administrative component of the platform.
VOXS is designed to meet the requirements of sensitive training environments. No audio is recorded, session data is ephemeral, student identity is not captured, and all communications are encrypted. For environments with specific accreditation, data residency, or information security requirements, please contact our enterprise team to discuss your specific compliance obligations.
Data residency options are available for enterprise deployments. If your organisation requires that data processing occurs within a specific region or jurisdiction, please speak to our enterprise team. We can discuss available options and what is technically feasible for your deployment.
Sign in to your account, click "New Session" from your dashboard, choose your speaking language, set a session PIN, and click "Create Session". A QR code will appear immediately — project this to your room. When your audience has joined, click "Start Speaking" and the translation begins.
For the best experience, use Google Chrome or Microsoft Edge on a laptop or desktop. These browsers support the Web Speech API used for real-time speech recognition. Safari and Firefox have limited speech recognition support and are not recommended for the trainer role.
If speech recognition stops — indicated by a red status badge — click the badge to restart it. This can occur after extended periods of silence. The system will attempt to restart automatically, but you can trigger it manually at any time. Ensure your browser has microphone permission granted before starting a session.
Yes. Click "Pause" during a live session. Students will see a paused status indicator. Click "Resume" to continue; speech recognition restarts automatically. Use this for breaks, Q&A periods, or when switching between topics.
Scan the QR code displayed in the training room using your phone's camera. This opens a web page in your browser — no application download is required. Select your language from the list and tap "Join Session". Translated text will appear on your screen as the trainer speaks.
No. Students do not need an account, do not submit any personal information, and do not need to install anything. Scan the QR code, choose your language, and you are connected. The platform collects nothing about you.
Yes. The translation screen includes A− and A+ controls to decrease or increase the text size. A dark mode toggle is also available — use whichever is more comfortable for your environment and vision.
The application will automatically attempt to reconnect. You will see a "Reconnecting" status. Once your connection is restored, translations will resume immediately. Content delivered during the disconnection period will not be shown retrospectively, but all new content will appear in real time.
We currently support 20+ languages including Arabic, Hindi, Urdu, Bengali, Tagalog, Indonesian, French, Spanish, Mandarin Chinese, Portuguese, Nepali, Tamil, Thai, Vietnamese, Malay, Pashto, and more. New languages are added on a regular basis. Arabic is fully supported with native right-to-left rendering.
VOXS is powered by one of the most capable AI language models currently available. It delivers natural, contextually aware translations that handle industry terminology, regulatory language, and technical content far more effectively than traditional rule-based or statistical systems. For Arabic, Hindi, and other major languages, accuracy is excellent.
VOXS is priced on an enterprise basis, tailored to your organisation's usage profile, languages, and deployment requirements. Contact our enterprise team for a tailored proposal.
Yes. The platform is content-agnostic — it translates whatever the trainer says. It is particularly well suited to structured training content such as health and safety briefings, compliance programmes, technical inductions, and regulatory training, where precise language matters. The AI handles industry-specific terminology across all supported languages.