Echosy

Summary

Echosy is a privacy-first offline transcription tool for macOS. All processing happens on-device with no cloud upload. Captures system audio and mic simultaneously, transcribes in real-time, and generates AI summaries.

What is Echosy?

Echosy is an on-device speech-to-text app for macOS. It uses Qwen3-ASR, MLX Whisper, and other local AI models with GPU acceleration to transcribe audio in 99+ languages. Users can record meetings, lectures, or conversations, with real-time transcription and export to Markdown, TXT, SRT, VTT, DOCX, or PDF. Built-in Gemma 4 and Qwen 3.5 models enable summaries without API keys, ideal for sensitive content and offline environments.

Core Capabilities

• Real-time transcription: System audio + mic capture with timestamped text output
• Multiple ASR models: Qwen3-ASR (Cantonese, Hokkien, Shanghainese dialect support), MLX Whisper, Tiny/Base/Small/Medium/Large V3 Whisper
• Universal dictation: Hotkey activation, Editor Mode auto line breaks, voice-triggered Enter
• On-device AI summaries: Built-in Gemma 4, Qwen 3.5, or connect OpenAI, Gemini, Claude, Ollama APIs
• Batch file transcription: Drag & drop WAV, MP3, MP4, MOV, M4A for batch processing
• Session history: Calendar view, search, replay past recordings with full transcripts
• Vocabulary biasing: Custom keywords and terminology for improved accuracy
• Real-time translation: Auto-translate during transcription, processed locally

Pros

• Privacy guaranteed: Audio and transcripts processed entirely on-device, no cloud upload, suitable for sensitive meetings and confidential work
• Offline operation: No internet required after model download, works in air-gapped environments and during travel
• Multi-language support: 99+ languages, Qwen3-ASR excels at Chinese dialects (Cantonese, Hokkien, Shanghainese)
• Flexible configuration: Choose from lightweight Tiny Whisper (75 MB) to Large V3 (3 GB) based on resource needs
• Free to start: Free tier offers 15 min/session, Qwen3-ASR 0.6B model, 3 AI summaries and chats per day

Positioning & best fit

Echosy is built for macOS users who prioritize privacy and offline capability, ideal for individuals and teams needing on-device transcription and AI summaries.

• Journalists, researchers, students: Record interviews, lectures, meetings with on-device transcription and summaries to protect source privacy
• Content creators, podcasters: Batch transcribe audio and video files, export SRT and VTT subtitles
• Enterprise users: Sensitive meetings, legal consultations, medical records requiring offline processing (Enterprise license required)
• Developers: Access meeting transcripts via Claude Code Plugin, integrate into AI coding workflows