即答会议 · AskLive

中文 · English below

隐私政策

生效日期:2026 年 11 月 6 日

一句话版本:我们不收集你的任何数据,也不运营任何服务器。App 从不保存音频;转写文字、问答记录和你导入的资料都只存在你自己的设备上。只有你主动启用的云端语音识别或 AI 功能(问答、纪要、领域校对),才会用你自己的密钥把相关内容直接发给你自己选择的服务商。

我们收集什么

什么都不收集。本 App 没有账号系统、没有分析统计、没有广告、没有任何第三方追踪 SDK。开发者不运营任何服务器,也就不存在「上传到我们这里」的任何数据。

数据去向一览

数据何时离开你的设备去向
会议音频仅当你在设置里主动选择「Fish Audio 云端」识别(默认为本地识别,音频不出设备)Fish Audio,使用你自己的密钥
问题文本 + 近期转写片段检测到问题并生成答案时你配置的 LLM 服务商,使用你自己的密钥
定稿转写片段 + 会前填写的领域与术语仅当你为该场会议指定领域或术语、且已配置 LLM 时,后台逐段校对同上
知识库资料片段仅当某个问题需要知识库作答时,随问题发送检索命中的相关片段同上
转写全文 + 问答记录仅当你点击「生成会议纪要」时同上
音频文件从不——App 不保存任何录音——
知识库文档原文从不——始终留在设备本地——

语音识别

默认使用 Apple 的设备端语音识别:音频在你的设备上转成文字,不会离开设备,也不会被 App 存成文件。若你在设置中主动切换到「Fish Audio 云端」识别,音频会按停顿分段、从你的设备直接上传至 Fish Audio 进行识别,使用你自己填入的密钥。该处理受 Fish Audio 的隐私政策约束。你可以随时切回本地识别。

AI 问答与会议纪要

这些功能需要你在设置中填入你自己的 LLM API 密钥(支持 DeepSeek、通义千问、OpenAI 等 OpenAI 兼容服务)才能启用。启用后,下列内容会从你的设备直接发送给你所选择的服务商,使用你自己的账户:检测到的问题文本、用于理解上下文的近期转写片段(约数百字)、以及在需要知识库作答时检索命中的资料片段。若你为会议指定领域或术语,定稿转写片段、会前填写的领域与术语以及少量前文还会逐段发送给同一服务商进行校对;即使使用本地语音识别,这一步也需要联网。不指定领域且不填写术语时,不会发生这项额外发送。点击「生成会议纪要」时,会发送该场会议的转写全文与问答记录(过长时取首尾片段)。这些数据不经过开发者,我们也无法访问,具体处理受你所选服务商的隐私政策约束。

不填写密钥时,实时转写、问题检测与提词器仍可正常使用,不会有任何数据外发。

本地存储

会议记录(转写文字、问答、纪要)、知识库文档与其索引、发言稿,全部通过 SwiftData 保存在你设备的 App 沙盒内,不会同步到任何服务器。你可以在 App 内删除会议、文档与发言稿;卸载 App 会一并清除全部数据。

API 密钥的存储

你的 LLM 与 Fish Audio 密钥仅保存在本机的 iOS 钥匙串(Keychain)中,只用于直接调用你选择的服务商。你可以随时在设置中删除;卸载 App 也会一并清除。

数据留存与删除

由于我们不收集任何数据,也就没有需要留存或可供删除的服务器端数据。所有内容都只存在于你的设备上,由你完全掌控。你发送给第三方服务商的数据,其留存政策由该服务商决定。

儿童隐私

本 App 不面向 13 岁以下儿童设计。如上所述,本 App 对任何用户都不收集数据。

政策变更

若未来功能变化影响本政策,我们会更新本页面并修改顶部的生效日期。重大变更会在 App 更新说明中注明。

联系我们

任何疑问请联系:zhuolun66@gmail.com


Privacy Policy

Effective date: November 6, 2026

In one sentence: We collect no data about you and operate no servers. The app never saves audio; transcripts, Q&A records and the documents you import stay on your own device. Only cloud speech recognition or AI features you enable (Q&A, summaries, domain-based correction) send relevant content out, directly to the provider you chose, with your own key.

What we collect

Nothing. The app has no accounts, no analytics, no ads, and no third-party tracking SDKs. The developer operates no servers, so there is nowhere for your data to be "uploaded to us".

Where data goes

DataWhen it leaves your deviceDestination
Meeting audioOnly if you actively choose "Fish Audio cloud" recognition in Settings (on-device is the default and audio never leaves)Fish Audio, using your own key
Question text + recent transcript contextWhen a question is detected and answeredThe LLM provider you configured, using your own key
Finalized transcript segments + meeting domain and termsOnly when you set a domain or terms for that meeting and have configured an LLM; segments are corrected in the backgroundSame as above
Knowledge base snippetsOnly when a question needs the knowledge base — the retrieved snippets are sent with the questionSame as above
Full transcript + Q&A recordsOnly when you tap "Generate Summary"Same as above
Audio filesNever — the app saves no recordings—
Your imported documentsNever — they stay on the device—

Speech recognition

By default the app uses Apple's on-device speech recognition: audio is turned into text on your device, never leaves it, and is never written to a file by the app. If you actively switch to "Fish Audio cloud" recognition in Settings, audio is chunked at natural pauses and uploaded directly from your device to Fish Audio using the key you entered. That processing is governed by Fish Audio's privacy policy. You can switch back to on-device recognition at any time.

AI Q&A and meeting summaries

These features require your own LLM API key in Settings (DeepSeek, Qwen, OpenAI and any OpenAI-compatible service). Once enabled, the following is sent directly from your device to the provider you chose, under your own account: the detected question text, a few hundred characters of recent transcript for context, and — when the knowledge base is needed — the retrieved document snippets. If you set a meeting domain or terms, finalized transcript segments, those domain hints, and a small amount of preceding text are also sent to the same provider for background correction. This requires a network connection even when on-device speech recognition is selected. With no domain or terms, no extra correction requests are made. Tapping "Generate Summary" sends that meeting's full transcript and Q&A records (truncated head-and-tail if very long). This traffic never passes through the developer and we cannot access it; processing is governed by your chosen provider's policy.

Without a key, live transcription, question detection and the teleprompter still work, and nothing is sent anywhere.

Local storage

Meeting records (transcripts, Q&A, summaries), imported documents and their index, and your scripts are all stored via SwiftData inside the app's sandbox on your device, and are not synced to any server. You can delete meetings, documents and scripts inside the app; uninstalling removes everything.

API key storage

Your LLM and Fish Audio keys are stored only in the iOS Keychain on your device and are used solely to call the providers you selected. You can delete them in Settings at any time; uninstalling the app removes them as well.

Retention & deletion

Since we collect nothing, there is no server-side data to retain or delete. Everything lives on your device, under your control. Data you send to a third-party provider is retained according to that provider's own policy.

Children's privacy

This app is not directed at children under 13. As described above, it collects no data from any user.

Changes

If future features change this policy, we will update this page and the effective date above, and note material changes in the app's release notes.

Contact

Questions? Reach us at zhuolun66@gmail.com.