更新日期：2026 年 10 月 24 日（写明运营者；新增"账号信息规范"和"AI 助理和智能体"）
本协议是你和本服务的运营者 北京锁儿哥科技发展中心（下称"我们"）之间的协议。联系方式：App 里的「意见反馈」。
"语崽"目前是内部测试版本，仅供受邀测试人员使用。功能可能变化、中断或出错，测试期间的数据可能被清除。请不要用它传递紧急或重要信息（例如报警、救援、安全指令）。
你需要用本人手机号通过短信验证码登录。请保管好你的手机，不要把验证码告诉他人。你可以随时在"设置"中注销账号。
账号信息规范（《互联网用户账号信息管理规定》）：你的昵称、头像，以及你给群和团队起的名字，不得：
你新设或修改这些信息时，我们会先审核（自动审核，拿不准的由人工审核），不符合的不予使用，并告诉你原因和怎么改；我们也会不定期检查已经在用的账号信息，不符合的经人工确认后改回默认，并通知你。对审核结果有异议，可以通过「意见反馈」申请人工复核。多次或严重违反的，我们可以暂停或终止提供服务。
你不得利用本服务发送违法违规内容，不得骚扰、威胁、诈骗他人，不得群发广告，不得尝试攻击、破解或滥用本服务（例如批量注册、刷验证码、猜测配对号）。违反者账号可能被停用。
"助理"和智能体（例如"记账助手"）的回答由人工智能（大模型）生成，可能不准确、不完整或者过时，不构成法律、税务、财务、医疗、投资等专业意见；重要的事请自己核实，或者咨询专业人士。把事情交给智能体之前会先问你同意，同时写明它由谁提供；也可以在设置里的"已授权的智能体"查看。
你不得利用助理和智能体生成、传播违法和不良信息，不得用来侵害他人的合法权益（例如冒充他人、编造虚假信息）。违反的，我们可以提醒、限制功能、暂停或终止服务，保存有关记录，并依法向主管部门报告。
AI 生成内容的标识（《人工智能生成合成内容标识办法》、GB 45438-2025）：
你把 AI 生成的内容发到别处（尤其是公开发布）时，应当按那里的规定主动声明并使用标识功能；不得恶意删除、篡改、伪造或者隐匿上述标识。对 AI 生成的内容有意见，点这条消息下面的"举报"（安卓上长按这条消息），或者通过「意见反馈」告诉我们。
你发出的语音、文字和图片属于你。你授权我们为了提供服务而存储并转发给你选择的好友、群或助理。保存多久见《隐私政策》第 5 节（例如语音保存 30 天后自动删除）。
测试版按"现状"提供，我们不保证服务不中断、不出错，也不保证消息一定送达。在法律允许的范围内，我们不对因使用测试版造成的间接损失承担责任。
我们如何处理你的个人信息，见《隐私政策》。
协议更新后会在本页面公布并更新日期，重要变化会在 App 里另行告诉你；继续使用即表示你同意更新后的协议。
Last updated: 24 October 2026 (who runs the Service; rules for account information; the AI assistant and agents)
These terms are between you and the operator of the Service, 北京锁儿哥科技发展中心 ("we"). Contact: Feedback in the app.
"Yuzai" is an internal test version for invited testers only. Features may change, stop or fail, and test data may be wiped. Do not rely on it for urgent or critical communication (emergency calls, rescue, safety instructions).
You log in with your own phone number and an SMS code. Keep your phone safe and never share the code. You can delete your account at any time in Settings.
Account information (PRC rules on internet user account information): your nickname and profile picture, and the names you give groups and teams, must not contain illegal or harmful content; must not pretend to be a government, military or party body, a company or organisation, a news outlet, a country or international organisation, or this platform's officials, support or admins, nor use another person's (including a public figure's) name or likeness; must not carry QR codes, web addresses, email addresses, phone numbers, WeChat IDs or other contact details (also when disguised) or advertising; and must not be misleading. New or changed account information is checked first (automatically, by a person when unclear); what doesn't comply isn't used, and we tell you why and what to change. We also check account information already in use from time to time; what doesn't comply is changed back to a default after a person confirms it, and you are told. You can ask for a person to look again through Feedback. Repeated or serious breaches may lead to suspension or termination.
Do not use the Service for illegal content, harassment, threats, fraud or spam, and do not attack, reverse-engineer or abuse it (mass sign-ups, requesting codes in bulk, guessing pairing codes). Accounts that break these rules may be suspended.
Answers from the assistant and from agents (such as the Bookkeeper) are generated by artificial intelligence (large language models). They may be inaccurate, incomplete or out of date and are not legal, tax, financial, medical, investment or other professional advice; check anything important yourself or ask a professional. Before anything is handed to an agent you are asked to agree, and told who provides it; you can also see this under "Agents you allowed" in Settings.
Do not use the assistant or agents to generate or spread illegal or harmful content, or to infringe on others' rights (for example impersonation or fabricated information). If you do, we may warn you, limit features, suspend or terminate the service, keep the relevant records and report to the authorities as required by law.
How AI-generated content is labelled (PRC Measures for Labelling AI-Generated Synthetic Content; GB 45438-2025):
When you post AI-generated content elsewhere (especially publicly), declare it and use the labelling tools there as its rules require; do not maliciously remove, alter, forge or hide these labels. To complain about AI-generated content, click Report under the message (on Android, long-press it), or tell us through Feedback.
Your voice messages, texts and pictures belong to you. You allow us to store them and deliver them to the friends, groups or assistant you choose. How long they are kept is in section 5 of the Privacy Policy (voice messages, for example, are deleted after 30 days).
The test version is provided "as is". We don't promise uninterrupted or error-free service or that every message arrives. To the extent the law allows, we are not liable for indirect losses from using it.
See the Privacy Policy for how we handle personal information.
Updates are published on this page with a new date, and important ones are also announced in the app; continuing to use the Service means you accept them.