这是什么
即答会议实时把会议语音转成文字,同时盯住每一个提问:能回答的立刻流式作答,回答不了的自动检索你导入的内部资料并标注来源,并把问题去掉口水话归档。会后一键生成纪要并导出。
AskLive transcribes your meeting in real time while watching for every question — streaming an answer to the ones it can handle, searching your own imported documents for the ones it can't, and filing each question away with the filler words stripped out.
常见问题 · FAQ
- 问题卡片显示「未配置 LLM API」怎么办?
- 问答与纪要功能需要你自己的 LLM API 密钥:设置 → LLM Provider → 添加 Provider,内置 DeepSeek、通义千问、OpenAI 预设,也可填任意 OpenAI 兼容服务的地址与模型名。填好后可点「测试连接」验证。不填密钥仍可正常使用实时转写、问题检测与提词器。Q&A and summaries need your own LLM API key (Settings → LLM Provider). Transcription, question detection and the teleprompter work without one.
- 英文会议识别得很差?
- 本地识别模型是分语种的,用中文模型识别英文会严重出错。纯英文会议请到 设置 → 语音识别 → 识别语言,选 English(首次使用该语言需下载模型);或直接切换到云端识别,云端会自动判别语言。On-device models are per-language — pick English under Settings → Speech Recognition → Language, or switch to cloud recognition which detects the language automatically.
- 本地识别和云端识别怎么选?
- 本地识别完全离线、零延迟,音频不出设备,适合涉密场合;Fish Audio 云端识别按停顿分段上传,对专业术语和标点更准、自动判别语言,但需要网络且音频会发送至 Fish Audio。两者随时可切。On-device is offline and private; Fish Audio cloud is more accurate on terminology and punctuation but uploads audio.
- 会议录音会被保存吗?
- 不会。App 不保存任何音频文件,只保存转写出来的文字、问答记录与纪要,全部存在你的设备本地。No audio is ever saved — only the text transcript, Q&A records and summary, all stored locally on your device.
- 导入的知识库资料会被上传吗?
- 资料本身始终留在设备本地,检索也在本机完成。只有当某个问题需要知识库作答时,检索命中的相关片段才会随问题一起发给你自己配置的 LLM 服务商。详见隐私政策。Documents stay on-device; only the retrieved snippets are sent with your question to the LLM provider you configured. See the privacy policy.
- 有问题被漏掉了怎么办?
- App 有四层检测(提前检测、规则触发、AI 确认、兜底扫描),后两层可在 设置 → 问题检测 中开关。若仍有遗漏,可在会议中点右上角「手动提问」,把最近说的内容直接交给 AI 提取并回答。Four detection layers, two of them toggleable in Settings. If something still slips through, tap "Ask" during the meeting.
- 提词器怎么用?
- 先在首页「发言稿」里粘贴或导入稿件,会议中点顶部文档图标选稿即可。浮层只盖住转写区,问答卡片照常可见。它会用实时转写定位你念到哪一行并自动滚动高亮;脱稿时停在原地等你回来。Add a script from the home screen, then tap the document icon during a meeting. It follows your voice and highlights the current line.
- App 需要哪些权限?
- 麦克风与语音识别,二者都是实时转写所必需。Microphone and Speech Recognition, both required for live transcription.
联系 · Contact
问题与建议:zhuolun66@gmail.com