← Package directory
Available on winget

Install XimilalaXiang.DeLive

Real-time Windows system-audio transcription with floating on-screen captions. Multi-ASR providers (Soniox, Volcengine), export TXT/SRT, history & auto-update.

Install with winget
winget install --id XimilalaXiang.DeLive
Upgrade
winget upgrade --id XimilalaXiang.DeLive
Uninstall
winget uninstall --id XimilalaXiang.DeLive

About XimilalaXiang.DeLive

Your last resort when all other paths are blocked. When subtitle export plugins fail, when platforms prevent downloads, when live streams have no captions, and when content is protected by DRM — system-level audio capture is your ultimate backup solution. Need to export subtitles or transcribed content for building knowledge bases, analysis, research, or any other purpose, but the platform restricts access? DeLive captures system audio and delivers clean, exportable text you own. 🎯 Core Features - 🎧 System-Level Au...

What's new in 2.2.3

Fixed / 修复 - 🔧 Provider switch diarization content loss — switching to a diarization-enabled provider (e.g. Soniox) no longer hides previously transcribed text; historical content now renders above speaker segments - 🔧 切换发言人识别模式内容丢失 — 切换到启用多发言人识别的提供商(如 Soniox)时,之前的转录内容不再消失,历史文本现在渲染在说话人分段之上 - 🔧 IndexedDB retry on open failure — if IndexedDB fails to open, subsequent calls now retry instead of permanently returning a rejected promise - 🔧 IndexedDB 打开失败后可重试 — IndexedDB 打开失败时后续调用现在会重试,而不是永久返回被拒绝的 Promise - 🔧 WebSocket close code RangeError — proxy cores now sanitize upstream close codes to valid ranges before forwarding, preventing RangeError crashes - 🔧 WebSocket 关闭码 RangeError — 代理核心现在在转发前将上游关闭码清理为合法范围,防止 RangeError 崩溃 - 🔧 S3 backup listing pagination — S3 backup list now fetches all objects using pagination instead of being capped at 100 - 🔧 S3 备份列表分页 — S3 备份列表现在使用分页获取所有对象,不再硬编码 100 条上限 - 🔧 Backup import overwrites provider API keys — importing backup data now preserves existing provider API keys and credentials instead of overwriting them - 🔧 备份导入覆盖 API Key — 导入备份数据时现在保留现有的提供商 API Key 和凭据,不再覆盖 - 🔧 AI post-processing timeout — non-streaming AI fetch calls (briefing, Q&A, mind map) now have a 120-second timeout to prevent indefinite hangs - 🔧 AI 后处理超时 — 非流式 AI 请求(摘要、问答、思维导图)现在有 120 秒超时,防止无限挂起 - 🔧 Local runtime exit handler race condition — fixed state race where the exit handler could read stale status by reordering spawn and state initialization - 🔧 本地运行时退出处理竞态 — 修复退出处理器读取陈旧状态的竞态条件,重新排列 spawn 和状态初始化顺序 - 🔧 Frontend TypeScript compilation — excluded Node-only proxy core files...

Read release notes

Version history

Version Updated Notes
2.2.3 Unknown Fixed / 修复 - 🔧 Provider switch diarization content loss — switching to a diarization-enabled provider (e.g. Soniox) no longer hides previously transcribed text; historical content now renders above speaker segments - 🔧 切...
2.2.2 Unknown Added / 新增 - 🎨 New color themes: Pink, Slate, Orange — three new accent color themes for more personalization options - 🎨 新增主题配色:樱粉、石板灰、橙韵 — 三个全新主题色,提供更多个性化选择 Changed / 变更 - 🎨 Default theme changed to Violet — new instal...
2.2.1 Unknown Added / 新增 - 🤖 AI side panel resizing — drag the left edge of the AI side panel to resize its width; width is persisted across sessions - 🤖 AI 侧边栏拉伸 — 拖拽 AI 侧边栏左边缘可调整宽度,宽度跨会话持久保存 - 🎙️ Real-time audio waveform — live wave...
2.2.0 Unknown Added / 新增 - 📁 File transcription — upload audio/video files and transcribe offline with ten cloud ASR engines: Soniox, Volcengine, ElevenLabs, Mistral AI, Gladia, Deepgram, AssemblyAI, Cloudflare Workers AI, SiliconFlow...
2.1.1 Unknown Fixed / 修复 - 🔧 Review tab scrolling — fixed transcript and all other review tabs (AI Correction, Overview, AI Analysis, Chat, Mind Map) being unable to scroll when content exceeds visible area; root cause was a missing f...
2.1.0 Unknown Added / 新增 - ☁️ Cloudflare Workers AI provider — new ASR provider based on Whisper, with low cost and free tier - ☁️ Cloudflare Workers AI 提供商 — 基于 Whisper 的新 ASR 提供商,价格低廉且有免费额度 - 🤖 AI post-processing smart text-source s...
2.0.1 Unknown Fixed / 修复 - 🌐 Comprehensive i18n overhaul — replaced 150+ hardcoded Chinese strings across 21 files with proper i18n keys - 🌐 全面 i18n 修复 — 21 个文件中 150+ 处硬编码中文替换为 i18n key - 🌐 Provider selector now displays translated na...
2.0.0 Unknown 🎨 Major UI/UX Overhaul / 重大界面重构 This release consolidates all beta changes (beta.1 through beta.10) into the stable 2.0.0 release. 此版本将所有测试版更改(beta.1 至 beta.10)整合为稳定的 2.0.0 正式版。 Added / 新增 - 🧭 Persistent left sidebar nav...
1.7.1 Unknown Added / 新增 - ☁️ S3-compatible cloud backup — upload, download, list, and delete session backups to any S3-compatible storage (AWS S3, MinIO, Cloudflare R2, etc.) - ☁️ S3 兼容云端备份 — 支持将会话备份上传、下载、列出和删除到任意 S3 兼容存储(AWS S3、MinI...
1.7.0 Unknown Added / 新增 - 🌐 Open API ecosystem — local REST API with 8 endpoints (/api/v1/sessions, /status, /topics, /tags, etc.) allowing external programs to access DeLive data - 🌐 开放 API 生态 — 本地 REST API 提供 8 个端点,允许外部程序访问 DeLive...
1.6.1 Unknown Fixed / 修复 - 🐛 Fixed topic assignment being silently discarded during session persistence — normalizeTranscriptSession() was missing the topicId field, causing it to be stripped every time a session was cached or saved -...
1.6.0 Unknown Added / 新增 - 📂 Topics feature — organize sessions into project-like topic containers (similar to ChatGPT Projects) with full CRUD, card-grid view, and per-topic detail pages - 📂 主题功能 — 将录制会话按主题/项目归类(类似 ChatGPT 项目),支持完整 C...
1.5.1 Unknown Added / 新增 - 🧠 AI Chat now renders Markdown responses with full GFM support: headings, lists, tables, bold/italic, and syntax-highlighted code blocks with one-click copy - 🧠 AI 对话现已支持 Markdown 渲染:标题、列表、表格、粗体/斜体,以及带语法高亮和一...
1.5.0 Unknown Added / 新增 - 🎨 Built a shared UI component library: Button, Badge, Switch, EmptyState, StatusIndicator, and DialogShell for consistent styling across the app - 🎨 构建共享 UI 组件库:Button、Badge、Switch、EmptyState、StatusIndicator...
1.4.4 Unknown Added / 新增 - 💬 Added session-level AI Q&A with persistent ask history, transcript-grounded answers, citation snippets, and support for multiple conversation threads inside a single saved session - 💬 新增会话级 AI 问答,支持持久化问答历史...
1.4.3 Unknown Added / 新增 - 🤖 Added AI post-process settings for OpenAI-compatible briefing generation, including configurable base URL, model, optional API key, and output language - 🤖 新增 AI 后处理设置,支持配置 OpenAI-compatible briefing 生成的 B...
1.4.2 Unknown Added / 新增 - 🌐 Added Soniox bilingual caption flow across transcript, preview, overlay, and session persistence - 🌐 新增 Soniox 双语字幕流,贯通主转录区、预览弹窗、悬浮字幕与会话持久化 - 🗣️ Added Soniox speaker diarization UI with speaker-grouped tra...
1.4.1 Unknown Added / 新增 - 🧪 Added rolling-audio and caption-wrap tests, increasing frontend coverage from 129 to 146 tests - 🧪 新增滚动音频缓冲与字幕换行测试,前端测试总数从 129 提升到 146 Changed / 变更 - ✅ Added frontend quality gates: runnable ESLint v9 conf...
1.4.0 Unknown Added / 新增 - 🏗️ Split monolithic transcriptStore into uiStore, settingsStore, sessionStore, tagStore with backward-compatible facade - 🏗️ 将单体 transcriptStore 拆分为 uiStore、settingsStore、sessionStore、tagStore,保留向后兼容外观层 - 🔬...
1.3.5 Unknown Changed / 变更 - Refactored electron/main.ts into focused modules for proxy, updater, caption window, tray, shortcuts, desktop source, app IPC, and local runtime orchestration - 将 electron/main.ts 进一步拆分为更聚焦的模块:代理、更新、字幕窗口、托...