# XVos (BoxVos) - full documentation for AI agents

Canonical site: https://www.boxvos.com/ (boxvos.com redirects/serves the same app)
Short index: https://www.boxvos.com/llms.txt
Last updated: 2026-10-17

> XVos is a browser-based AI voice intercom and real-time cross-language translation web app. No dedicated hardware and no app download are required.

Scope note: this document only restates what the XVos site and app state about themselves. Where something is not published, it says "not published". Do not infer facts that are not written here.

## 1. What XVos is

- Name: XVos (also written BoxVos / XVox AI in some places). Domain: boxvos.com.
- Tagline on the home screen (Chinese): 无界对讲 即时传声 ("intercom without borders, instant voice").
- Subtitle on the home screen (Chinese): AI 边缘全双工实时语音对讲与即时互译系统 ("AI edge full-duplex real-time voice intercom and instant mutual translation system").
- Category: communication web application (schema.org WebApplication), free to open in a browser; usage minutes are topped up inside the app (see section 6).
- Creator: indie developer @indie_maker_fox (https://x.com/indie_maker_fox).
- Type of product: a web app opened in a normal phone, tablet or computer browser. The app configuration also describes camera and microphone permissions (camera for QR scanning, microphone for voice).

## 2. Features (as stated on the site)

- Zero hardware, zero install: works in a regular browser on phone, tablet or PC ("零硬件·免安装 随时即时对讲"; "无需专用设备，浏览器扫码直接通话").
- Room-based voice intercom: one user creates a room (创建房间); others join by scanning the room's QR code (扫码加入) or typing the room code (输入房间号).
- Full-duplex real-time voice: speakers can talk and interrupt naturally instead of strict turn taking.
- Real-time cross-language translation: each participant picks their own language and hears the others translated; multi-party rooms are supported (the 30-minute pack lists "3 or more people in real-time calls with mutual translation").
- Language selector in the home screen: 中文 (Chinese), English, 日本語 (Japanese), 한국어 (Korean), Spanish, French, German. The site additionally claims support for "100+ languages" (语种 100+).
- Claimed latency: under 100 ms ("< 100ms 极低延迟全双工"; edge neural voice streaming).
- Claimed audio quality: 24 kHz HD natural voice ("24kHz HD"), with adaptive noise reduction that preserves the speaker's tone.
- Claimed security/privacy: end-to-end encrypted direct (peer-to-peer WebRTC) streaming; the site says no audio recording is retained ("端到端安全与零留存隐私").
- PC view: a desktop layout with a marketing panel on the left and a phone-style app frame on the right; the PC page also shows a QR code so a phone can open the same app.
- Countdown reminders: the app announces remaining minutes by voice at 5 minutes and 1 minute (stated in the 30-minute pack description).
- Sign-in: creating a room requires signing in (登录). The app keeps a per-user remaining-minutes balance.
- Mobile app: pack descriptions mention a phone-app "local AI, no data traffic" mode and offline/cloud model switching; availability of store apps is not published.

## 3. How to use XVos

1. Open https://www.boxvos.com/ in a browser (phone or PC).
2. Choose your language in the language selector (default 中文).
3. To start a conversation: sign in, then press 创建房间 (Create room). Share the room code or QR code with the other people.
4. To join: press the scan icon to scan a QR code, or type the room code into 输入房间号 and press the arrow.
5. Talk. Each participant hears the others in their own chosen language.
6. When minutes run low, top up from the account/membership area (see section 6).

## 4. Home screen layout (for agents driving the UI)

- Top: XVos logo with "VOICE AI OS", and a 登录 (Sign in) button.
- Centre: headline 无界对讲 / 即时传声, the subtitle, and a control bar with 创建房间, a scan (fullscreen-style) icon, the room-code input with an arrow button, the language selector, and a user/login icon.
- Bottom: four stat tiles (延迟 <100ms, 音质 24kHz HD, 安全 E2E, 语种 100+) and footer links: 退款政策 (Refund Policy), 联系我们 (Contact Us), 隐私政策 (Privacy Policy), 服务条款 (Terms of Service), and a link to @indie_maker_fox on X.
- The legal pages (refund, contact, privacy, terms) open inside the app; they do not have separate public URLs.

## 5. Languages

- Selectable in the UI today: Chinese, English, Japanese, Korean, Spanish, French, German.
- Site claim: "100+ languages". A complete list of the 100+ languages is not published.
- Interface text exists in Chinese, English and Japanese.

## 6. Pricing (as shown in the in-app top-up dialog)

There is no separate public pricing page. The in-app top-up dialog ("XVos 时长充值与会员中心") shows minute packs, billed at about 0.2 CNY per minute. Minutes are deducted on use and do not expire ("分钟数按需扣除不过期").

| Pack | Minutes | Price (CNY) | Price (USDT) | Price (HKD) |
|---|---|---|---|---|
| 10分钟体验包 (10-minute trial) | 10 | 2.00 | 0.28 | 2.2 |
| 30分钟通勤包 (30-minute commuter pack, marked "most popular") | 30 | 6.00 | 0.85 | 6.6 |
| 60分钟商务包 (60-minute business pack) | 60 | 12.00 | 1.70 | 13.2 |
| 100分钟长聊加油包 (100-minute long-chat pack) | 100 | 19.90 | 2.80 | 22.0 |

- Payment options shown in the dialog: Binance Pay / USDT, Hong Kong ZA Bank (FPS), and redeeming a voucher/redeem code (兑换码). Offline transfers are reviewed manually before minutes are credited.
- Pricing may change and the top-up dialog in the app is the source of truth. No subscription plans or enterprise pricing are published.

## 7. Refunds, privacy and terms

- The app has Refund Policy, Privacy Policy and Terms of Service pages. The Refund Policy says problems that prevent use of purchased minutes (for example server downtime) can be reported, and pro-rated refunds or extra minutes may be provided; refund requests are sent by email to the contact address below with the account email, transaction ID or purchase date, and a short reason.
- Privacy claims on the site: encrypted peer-to-peer WebRTC streaming; no audio recording retained.

## 8. Contact

- Contact: xiebaoxin@mail.grokbot.com
- Creator on X: https://x.com/indie_maker_fox
- Company name, address, phone number and public API are not published.

## 9. Machine-readable resources

- https://www.boxvos.com/llms.txt
- https://www.boxvos.com/llms-full.txt (this file)
- https://www.boxvos.com/index.md
- https://www.boxvos.com/about.md
- https://www.boxvos.com/sitemap.xml
- https://www.boxvos.com/robots.txt
- Structured data (JSON-LD, schema.org WebApplication and FAQPage) is embedded in the HTML of https://www.boxvos.com/

## 10. FAQ (from the site's structured data)

Q: Does XVos need extra hardware or a dedicated intercom?
A: No. A normal phone, tablet or computer browser is enough.

Q: Do I need to install an app?
A: No. Open the web link or scan the room's QR code with the phone camera to join a room.

Q: How does XVos translate across languages in real time?
A: It uses low-latency audio streaming with an edge AI neural translation engine for two-way real-time interpretation, so people with different native languages can talk naturally.