This is Ryusei__46 speaking.
In this article, I'll introduce some recommended add-ons for Windows' open-source screen reader "NVDA" specifically for low-vision users.
Since NVDA was originally developed for completely blind users, I found that many of its features can be somewhat cumbersome for low-vision individuals to use.
As a low-vision user myself, I'd like to share the NVDA add-ons I regularly use as a reference. Additionally, I'll also mention some high-quality free speech engines available for use.
This article will be updated incrementally as I discover more useful add-ons. If you know of any particularly good ones, please share them in the comments section.
Add-on "NoBeepsSpeechMode"
This add-on modifies the default behavior when pressing "NVDA key (Unconvert)" + "S". Normally, this sequence performs: "beep-based audio feedback" ⇒ "pause speech output" ⇒ "resume speech output". However, I've never actually used the "beep-based feedback" feature, and finding myself needing to press the shortcut twice to pause speech was quite cumbersome.
After installing this add-on, the "beep-based audio feedback" feature is removed, allowing you to toggle between pausing and resuming speech output with a single shortcut key press.
This add-on becomes active immediately after installation without any additional configuration required.
Add-on "Focus Highlight"
This add-on visually highlights navigator objects that have focus by drawing a border around them. While NVDA's built-in "Vision" feature provides similar functionality, it lacks the ability to change the border color—this add-on allows you to customize the color to your preferred visibility setting.
The above is a feature I've tested in my blog article - when your mouse pointer hovers over text, it appears framed in light green. The border color can be freely customized through the addon's settings.
This feature proves incredibly useful - it clearly shows you the current focus range, even when dealing with paragraphs containing numerous text links or embedded HTML tags, as it visually outlines the active focus area. This makes it much easier to determine where to next move your mouse for focus.
To configure these settings, go to the "NVDA Settings" screen and select "Focus Highlight" under the "Categories" section. Here, you can customize the border color, width, and shape for three different states: the "Focus Mode" and "Browsing Order" used in browsers, and the "Navigator Objects" used on Windows.
Since color values must be entered in hexadecimal format, we recommend selecting your preferred color from sites like the following and copying/pasting the code:
色の名前とカラーコードが一目でわかるWEB色見本
● Enable/Disable Add-ons
You can toggle the Focus Highlight add-on on or off by pressing: NVDA Key (Unconverted) + Alt + P.
NVDA SpeechDic Creator
While this isn't officially distributed as an NVDA add-on, it serves as a convenient tool for creating, editing, and sharing NVDA's speech dictionary.
This tool is developed using VB Script and remains extremely lightweight. Due to the nature of VB Script, its user interface consists of a simple message box-style window with step-by-step operations, making it easy for anyone to use.
NVDA already includes built-in functionality for creating and editing dictionaries, but it doesn't support importing or exporting dictionary data. This creates inconvenience when transitioning environments or wanting to share dictionary data with other users.
This is where the tool comes in handy. Detailed usage instructions are provided in the readme file included with the tool, so please refer to that for guidance.
NVDAスピーチディッククリエイター
Bonus: The paid voice engine "KC-Talker"
This is a voice engine add-on that incorporates "Voice Text," a high-performance paid text-to-speech engine developed by HOYA. I had previously used HOYA's voice engine when working with "PC-Talker" screen readers. The quality is absolutely top-notch—so excellent that I found no room for improvement.
That said, since the add-on costs just under ¥20,000, I recommend considering a purchase if you're financially able.