Echosy is a private, on-device audio transcription and dictation tool designed for macOS. It allows users to record both system audio and microphone input without any internet connection, ensuring complete privacy. The app uses local AI models such as Whisper and Qwen3-ASR to transcribe audio in real-time, summarize content, and even dictate text directly into any application. Since all processing happens on the user's Mac, no data leaves the device, making it ideal for sensitive or confidential use cases.
With support for a wide range of audio formats and multiple ASR models, Echosy provides flexible and powerful transcription capabilities. Users can also enhance transcripts with auto-punctuation, translation, and custom prompts. AI summaries are generated using OpenAI, Gemini, Ollama, or other compatible APIs, and all features run locally without cloud dependency. Whether for personal notes, professional meetings, or batch file transcription, Echosy delivers a secure and efficient audio intelligence solution.
Echosy operates entirely on the user's Mac, utilizing macOS ScreenCaptureKit to capture audio from any application. Once recorded, the audio is processed by an on-device AI model (such as Qwen3-ASR) to generate a real-time transcript. The transcript can be enhanced with auto-punctuation, translation, and custom prompts. AI summaries are created using compatible LLMs, and the entire process remains private, with no data sent to external servers.
Users can dictate text anywhere on macOS using a customizable hotkey, and transcripts can be exported in various formats like MD, TXT, SRT, VTT, DOCX, and PDF. For batch processing, files can be dragged and dropped for offline transcription. All sessions are stored in a session history for easy access and replay.
