Capture system audio. Transcribe with six ASR backends. Review with AI — summaries, chat, mind maps. All local-first.
Six ASR engines, AI-powered review, floating captions — all running locally.
Soniox (real-time streaming with translation & diarization), Volcengine, Groq, SiliconFlow, OpenAI-compatible, and whisper.cpp. Three modes.
IndexedDB storage, Electron safeStorage encryption, context isolation. Your data stays.
Summary, action items, keywords, chapters. Multi-thread chat. Mind maps via Markmap.
Always-on-top draggable subtitle window. Source, translated, bilingual modes. Fully customizable font, color, shadow and background.
REST API, WebSocket streaming, MCP server for Claude Desktop and Cursor.
Cyan, Violet, Rose, Green, and Amber — light and dark modes. Command Palette.
WebDAV + S3-compatible. Nutstore, Nextcloud, Alist, Amazon S3, MinIO. Auto-create directories.
From live transcription to AI-powered insights.
Capture any system audio — meetings, lectures, podcasts. 60+ languages with speaker detection.
AI summaries, action items, mind maps. Chat with transcripts.
Ask questions, get clarifications, extract info — through a natural chat interface with Markdown rendering.
Six ASR providers, any OpenAI-compatible AI model, and MCP integration.
API Endpoints
Languages
Tests Passing
Free and open source. Windows, macOS, Linux.
Apache-2.0 License. All releases →
DeLive is fully open source. Star, report, or contribute.