English | 中文 | 日本語

OpenEcho

Speak once. Insert clean text anywhere.

OpenEcho is a macOS desktop app that turns your voice into polished text, translations, answers, and structured memos — all from a single shortcut. Speech recognition runs locally on your machine, your audio never leaves your device.

---

Quick Start

1. Download the latest .dmg from the Releases page (Apple Silicon only)
2. Install — open the .dmg, drag OpenEcho to Applications
3. Launch — the setup wizard walks you through permissions, microphone test, and model download
4. Use — press the global shortcut, speak, done. Text appears wherever your cursor is.

If macOS blocks the app, run: sudo xattr -dr com.apple.quarantine /Applications/openecho.app

---

Why OpenEcho?

Most voice tools send your audio to the cloud, lock you into one use case, or break your workflow. OpenEcho is different:

• Privacy first — Whisper runs locally. Your voice stays on your Mac.
• One shortcut, five modes — Input, Translate, Question, Memo, Voice Generate. No app switching.
• Works with any app — Text is pasted directly into whatever you're working in.
• LLM is optional — Core voice-to-text works offline. Connect an OpenAI-compatible API for extra features when you want them.

---

Screenshots

Setup

Guided first-launch wizard — permissions, microphone test, model download, and optional LLM configuration.

Voice Input

Press a shortcut, speak, and polished text is pasted wherever your cursor is.

Dashboard

Overview of your settings, recent transcriptions, and quick access to all features.

History

Browse, play back, copy, or delete past transcriptions. Audio recordings are saved alongside each entry.

Voice Generate

Type text and generate speech with local or remote TTS. Preview, save, and choose from multiple languages.

Settings

Configure LLM, TTS, recognition model, shortcut, theme, and more.

---

What can it do?

Input

Press the shortcut, speak, and the recognized text is cleaned up and pasted into your current app. Perfect for emails, messages, code comments — anywhere you type.

Translate

Speak in your language, get text in another. Choose your target language, and OpenEcho handles the rest through your LLM endpoint.

Question

Ask a question by voice. OpenEcho shows both your question and the AI-generated answer in a floating panel. Great for quick lookups without leaving your workflow.

Memo

Dictate a messy thought. OpenEcho organizes it into a structured memo you can review and edit before inserting.

Voice Generate

Type text and hear it spoken aloud. Supports local TTS (Kokoro, runs offline) or remote TTS APIs (OpenAI-compatible). Preview, save as audio file, and choose from multiple languages.

---

How it works

Global shortcut
      ↓
Microphone capture
      ↓
Local Whisper model
      ↓
  ┌─────────────────────────────────────────────────┐
  │  Input     → clean text → paste into your app   │
  │  Translate → LLM translation → paste            │
  │  Question  → LLM answer → floating Q&A panel    │
  │  Memo      → LLM organization → editable memo   │
  └─────────────────────────────────────────────────┘

---

Build from source

Requires macOS, Node.js, and Rust toolchain.

npm install
npm run tauri dev       # development
npm run tauri build     # production build

Release

Releases are automated via GitHub Actions. To create a new release:

git tag v0.1.0
git push origin v0.1.0

This will:

1. Build .dmg installer for Apple Silicon Mac
2. Generate a changelog from commit history since the last tag
3. Create a GitHub Release with the installers and changelog

Note: The app is not code-signed. On first launch, right-click → Open, or allow it in System Settings → Privacy & Security.

License

MIT