System Audio Capture | 12 ASR Providers | Local-First AI Review Workspace
🌐 Official Website · 📖 Documentation · 📋 Getting Started · ⬇️ Download
DeLive is a desktop transcription workspace for system audio. It captures whatever your computer is playing, routes the audio through any of twelve ASR backends, keeps everything on your machine, and turns completed transcripts into searchable history with a full AI Review Desk — AI transcript correction, rich Markdown-rendered chat, Q&A threads, structured briefings, and mind maps. It also supports uploading audio/video files for offline transcription, with ten cloud engines available for file transcription.
| Live Transcription | Caption Overlay | MCP Integration |
|---|---|---|
| Real-time transcription with 12 ASR providers | Draggable always-on-top floating caption window | External AI tools access DeLive via MCP protocol |
 |
 |
 |
| AI Overview | AI Correction | AI Chat |
|---|---|---|
| Summary, action items, keywords, and chapters | Quick Fix and Review & Fix modes with diff view | Multi-thread conversation with cited references |
 |
 |
 |
- System-audio capture — browser video, live streams, meetings, courses, podcasts, or any other playback source
- Twelve ASR backends — Soniox, Volcengine, Groq, SiliconFlow, Mistral AI, Deepgram, AssemblyAI, ElevenLabs, Gladia, Cloudflare Workers AI, OpenAI-compatible local services, and local whisper.cpp
- File transcription — upload audio/video files for offline transcription with ten cloud engines
- AI Review Desk — transcript correction (Quick Fix / Review & Fix), structured briefings, multi-thread chat, Q&A, and mind maps
- Floating caption overlay — always-on-top window with source / translated / dual display modes
- Soniox bilingual & speaker-aware — realtime translation, dual-line captions, speaker diarization
- Topics — organize sessions into project-like containers
- Local-first — sessions, tags, topics, and settings stored locally; optional S3/WebDAV cloud backup
- Open API & MCP — local REST API, real-time WebSocket, MCP server for AI agents
- Cross-platform — Windows, macOS, and Linux
📖 Full feature details in the documentation.
| Platform | Files |
|---|---|
| Windows | .exe installer, portable .exe |
| macOS | .dmg, .zip (Intel x64 and Apple Silicon arm64) |
| Linux | .AppImage, .deb |
| Provider | Type | Transport | File | Highlights |
|---|---|---|---|---|
| Soniox V4 | Cloud | Realtime streaming | Yes | Token-level transcription, realtime translation, bilingual captions, speaker diarization |
| Volcengine | Cloud | Realtime streaming | Yes | Chinese-oriented realtime path with embedded proxy |
| ElevenLabs | Cloud | Realtime streaming | Yes | Scribe v2 Realtime; 99 languages |
| Mistral AI | Cloud | Realtime streaming | Yes | Voxtral Realtime |
| Gladia | Cloud | Realtime streaming | Yes | Solaria-1; 100+ languages; <300ms latency |
| Deepgram | Cloud | Realtime streaming | Yes | Nova-3 / Nova-2 streaming |
| AssemblyAI | Cloud | Realtime streaming | Yes | Universal-3 Pro streaming |
| Cloudflare Workers AI | Cloud | Windowed batch | Yes | Whisper-based; low cost with free tier |
| SiliconFlow | Cloud | Windowed batch | Yes | SenseVoice, TeleSpeech, Qwen Omni |
| Groq | Cloud | Windowed batch | Yes | Whisper large-v3-turbo / large-v3 |
| Local OpenAI-compatible | Local | Windowed batch | — | Works with Ollama or compatible gateways |
| Local whisper.cpp | Local | Electron-managed | — | Fully local; DeLive manages binary and model lifecycle |
📖 Provider setup details: API Keys Guide · Provider Comparison
git clone https://github.com/XimilalaXiang/DeLive.git
cd DeLive
npm run install:all
npm run devgraph TB
subgraph "Desktop Shell"
EM[Electron Main Process]
WIN[Main Window]
CAP[Caption Overlay Window]
DESK[Tray / Shortcut / Auto Launch / Updater]
SEC[IPC Security / SafeStorage / Diagnostics]
end
subgraph "Renderer"
UI[React App]
STORES[Zustand Stores]
CFG[Provider and Runtime Setup]
PREV[History / Preview / AI Workspace]
end
subgraph "Orchestration"
ASR[useASR]
CAPMGR[CaptureManager]
PROVSESS[ProviderSessionManager]
CAPBR[CaptionBridge]
end
subgraph "Capture Pipeline"
GDM[getDisplayMedia]
MR[MediaRecorder<br/>WebM / Opus]
AP[AudioWorklet<br/>PCM16 16kHz]
end
subgraph "Provider Layer"
REG[Provider Registry]
SON[Soniox]
VOL[Volcengine]
ELB[ElevenLabs]
MIS[Mistral AI]
GLA[Gladia]
DPG[Deepgram]
AAI[AssemblyAI]
CFL[Cloudflare Workers AI]
SIL[SiliconFlow]
GRQ[Groq]
LOA[Local OpenAI-compatible]
WCP[whisper.cpp Runtime]
end
subgraph "Electron Services"
PROXY[Embedded Multi-Provider Proxy]
RTM[Local Runtime Controller]
end
subgraph "Persistence"
REPO[Session Repository]
IDB[IndexedDB]
LS[localStorage]
SAFE[safeStorage]
end
UI --> STORES
UI --> CFG
UI --> PREV
UI --> ASR
ASR --> CAPMGR
ASR --> PROVSESS
ASR --> CAPBR
CAPMGR --> GDM
GDM --> MR
GDM --> AP
PROVSESS --> REG
REG --> SON
REG --> VOL
REG --> ELB
REG --> MIS
REG --> GLA
REG --> DPG
REG --> AAI
REG --> CFL
REG --> SIL
REG --> GRQ
REG --> LOA
REG --> WCP
MR --> SON
MR --> LOA
AP --> VOL
AP --> ELB
AP --> MIS
AP --> GLA
AP --> DPG
AP --> AAI
AP --> CFL
AP --> SIL
AP --> GRQ
AP --> WCP
VOL --> PROXY
MIS --> PROXY
DPG --> PROXY
AAI --> PROXY
ELB --> PROXY
GLA --> PROXY
WCP --> RTM
STORES --> REPO
REPO --> IDB
REPO --> LS
CFG --> SAFE
UI --> EM
EM --> WIN
EM --> CAP
EM --> DESK
EM --> SEC
EM --> PROXY
EM --> RTM
CAPBR --> CAP
style UI fill:#61dafb,color:#000
style EM fill:#334155,color:#fff
style CAP fill:#f472b6,color:#000
style REG fill:#f59e0b,color:#000
style PROXY fill:#10b981,color:#fff
style RTM fill:#0f766e,color:#fff
style SEC fill:#ef4444,color:#fff
style SAFE fill:#a855f7,color:#fff
style IDB fill:#3b82f6,color:#fff
📖 Detailed architecture: Overview · Providers · Electron IPC · Data Flow · Security
DeLive/
├── electron/ # Electron main process, windows, tray, IPC, updater, runtime, Open API server
├── frontend/ # React renderer app, providers, stores, UI components, tests
├── shared/ # Shared TypeScript contracts and provider proxy helpers
├── server/ # Standalone proxy server for debugging
├── mcp/ # MCP server for AI agents (Claude, Cursor, etc.)
├── skills/ # Agent skill definitions
├── scripts/ # Icon generation, runtime staging, release notes
├── docs/ # VitePress documentation site source
├── landing/ # Landing page source
└── package.json
📖 Full project map: Project Structure
| Layer | Technology |
|---|---|
| Desktop app | Electron 40 |
| Frontend | React 18.3 + TypeScript 5.6 + Vite 6 |
| Styling | Tailwind CSS 3.4 |
| State management | Zustand 4.5 |
| Testing | Vitest 4 (314 tests / 32 files) |
| Persistence | IndexedDB, localStorage, Electron safeStorage |
| Packaging | electron-builder + GitHub Actions |
DeLive exposes a local REST API, real-time WebSocket, and an MCP server for AI agents — all disabled by default, with optional Bearer token authentication.
📖 Full API reference: REST · WebSocket · MCP Server · Authentication · Agent Skill
- System requirements: Windows 10+, macOS 13+, or Linux with PulseAudio loopback support.
- Provider proxies are embedded in Electron — no separate backend needed for desktop usage.
- Tray behavior: closing the main window hides to tray instead of exiting.
- Auto-update: supported on Windows, macOS, and Linux AppImage.
Windows may show a SmartScreen warning on first launch. Click More info → Run anyway.
Apache License 2.0
- BiBi-Keyboard for multi-provider architecture inspiration
- ByteDance — Volcengine speech recognition service and Lark AI Campus Challenge support
- LINUX.DO community — a place where we learned a great deal and received generous support
Made by XimilalaXiang