Home

VoxType Wiki - Windows Voice Typing and Speech-to-Text

VoxType is a lightweight Windows desktop voice typing, dictation, and speech-to-text tool. Put the cursor in any input box, press the global shortcut, speak, and VoxType will transcribe your voice with Doubao streaming ASR, optionally polish the result with an OpenAI-compatible LLM, copy it to the clipboard, paste it into the active field, and restore the previous clipboard when possible.

声写 VoxType 是一个 Windows 桌面语音输入、语音转文字和听写工具。把光标放到任意输入框后，按全局快捷键开始说话，VoxType 会完成录音、豆包流式 ASR、可选大模型润色、写入剪贴板、自动粘贴和剪贴板恢复。

Use Cases / 适合场景

• Windows voice typing and speech-to-text in chat apps, browsers, editors, forms, and office tools.
• 中文语音输入、英文听写、多语言语音转文字，以及需要实时字幕的桌面输入场景。
• Doubao ASR transcription with optional LLM polishing for cleaner long-form dictation.
• A local-first open-source workflow with conservative privacy defaults.

简体中文

• 用户配置指南
• 功能特性与使用优化
• 常见问题与排障

English

• User Configuration Guide
• Features and Usage Optimization
• Troubleshooting

Recommended Reading Order

1. Configure Doubao ASR first.
2. Add hotwords and optional prompt preferences.
3. Adjust daily preferences such as shortcut, paste mode, microphone, captions, startup, and tray behavior.
4. Open advanced settings only when troubleshooting latency, paste compatibility, ASR connection, LLM timeout, or diagnostics.

Current Defaults Worth Knowing

• Main trigger: Ctrl + Q.
• Right Alt and middle mouse triggers: off by default.
• Recent context and automatic hotword candidates: off by default.
• Local silence fallback: continuous low volume stops recording after 30 seconds by default, with a 0.03 threshold.
• Doubao server endpointing example value: end_window_size = 800.
• Update prompts provide an "Update now" action when a new installer is available.

Privacy Basics

• config.toml, logs, local context files, and usage stats are local files and should not be committed.
• Usage statistics record duration, character count, speed, and time estimates, not recognized text.
• Recent context and automatic hotword history are disabled by default.
• Diagnostic reports and logs should not include real API keys, recognized text, hotwords, prompts, recent context, or Windows username paths.

Search-Friendly FAQ

Is VoxType a Windows dictation app?

Yes. VoxType is a Windows desktop dictation app that turns microphone speech into text and pastes it into the active input field.

Does VoxType require Doubao ASR?

Yes for the main speech-to-text workflow. Doubao ASR App Key and Access Key are required before recording, recognition, and paste actions are unlocked.

Is LLM polishing required?

No. LLM polishing is optional. VoxType can run as a pure ASR voice input tool, and only calls an OpenAI-compatible LLM when polishing is enabled and configured.