Privacy & analytics

Anonymous visit analytics

本站使用自己的统计接口记录公开页面的访问量、匿名访客、设备类别、到站来源域名和入口页。服务器通过本地 IP 地域库估计国家、省州和城市，保存脱敏网段：IPv4 隐去末段，IPv6 只保留 /48 前缀。原始 IP 只用于本地查询，不写入统计库，也不发送给Third-party地域服务。浏览器保存一个Active期 30 天的随机 Cookie；数据库仅保存其哈希，用于所选时间段内去重。统计不记录实名身份、原始 IP、Search问题或来源网址中的路径与查询词。

When Do Not Track (DNT) or Global Privacy Control (GPC) is enabled, visit and selection, export, and favorite event analytics are disabled. You may clear site cookies; this affects anonymous deduplication and may disconnect private favorites associated with other cookies.

来源与免费工具操作

到站来源只保留域名；实际存在的 utm_source、utm_medium、utm_campaign、utm_content、utm_term 只接受限长活动标签。Search问题、密钥及包含邮箱的活动值不发送。筛选、估算、保存、配置和Official access点击按匿名浏览器计数，不证明真实人数、成交或返佣。

保存筛选和比较使用当前设备的本地存储，最多保留 20 组。items件只在你主动保存和恢复时使用；清除本地存储后无法恢复。网站不再按照访问者 IP 网段、浏览器特征或行为规则限制公开页面、目录、robots.txt、sitemap 或公开内容接口。旧版网段配额和行为规则已停用，相关后台数据只作历史记录，不再影响请求。Search和 AI 爬虫仍按其声明的 User-Agent 分类用于抓取统计；可获取官方 IP 范围时用于来源核验统计，不参与公开内容放行判断。服务器访问 IP 可能在本机用于匿名地域估计和必要的安全日志；分析统计仅保存脱敏网段。管理接口和需要身份验证的 API 继续按登录凭据、来源校验、业务频次和数据权限处理。

Model queries & usage

When task interpretation is enabled, your submitted request may be sent to the configured primary or backup provider. The model is used only for task classification, and the provider's data policy applies to that call. Rule-based interpretation does not send your text to a model provider.

Usage records include model, provider, purpose, primary or backup channel, time, status, duration, and token counts reported by the provider. Query text, responses, and API keys are not saved. Missing usage is marked unknown and is not estimated from text length.

Saved models & configuration exports

Private favorites and export history are linked to a separate anonymous browser credential and stored on the server. Export history may contain text you entered into a configuration example; do not enter real keys or sensitive data. You can delete records from Saved models. Administration uses a separate authenticated session.

Crawler & operations records

Search and AI requests are classified by their declared User-Agent. Analytics stores only the date, public path, agent name, and request count. Reverse DNS identity verification is not performed, so these records do not establish search indexing or AI citations. Network services may retain necessary security and diagnostic logs, separately from website analytics.

Analytics is available only in the protected admin area. Browser settings, blockers, and missing provider usage reports can make statistics incomplete.

Baidu Analytics and unified reporting

Public pages load Baidu Analytics (hm.baidu.com). It may receive page paths, referrers, device and browser information, and network addresses, and may use cookies. Baidu processes that data under its service terms. The script is not loaded when DNT or GPC is enabled or in the admin area.

Unified site reporting also uses server access logs, deduplicates by IP, and filters identified crawlers. This differs from browser analytics: disabling scripts or enabling DNT/GPC does not prevent necessary server logs. The unified dashboard receives only anonymous aggregates and masked network prefixes; full IP addresses do not cross servers.

本站另外自托管 Umami 作为独立对照统计，公开页面从 aiaiai.help/umami/ 加载脚本，并向独立数据库发送页面路径、来源及浏览器信息。独立的腾讯云服务器会接收到请求网络地址；Umami 用自身规则计算访客和会话，不与本站 Cookie 或服务器日志统计total用身份标识。开启 DNT 或 GPC 时不加载脚本，后台页面也不加载。新统计从启用时开始，历史数据无法回填；不同口径的数字应按相同日期范围比较。