Voice Master

The built-in Chrome speech engine is free, but not the most accurate: it can miss punctuation or even insert a wrong word that sounds similar. "Voice Master" can pass your dictated text through a modern AI model, which corrects spelling, punctuation, style — and replaces misrecognized words with what you most likely meant, judging by the context.

Open settings (the ⚙️ gear on the panel), choose your AI provider — OpenAI (ChatGPT), Anthropic (Claude) or DeepSeek — and paste your API key. Improve on demand: in the dictate window, press "✨ Improve current text" and the whole text is corrected in place (Ctrl+Z restores the original). NEW — Improve on the fly: enable the checkbox in settings, and every dictated phrase is sent to the AI and inserted already corrected. Phrases are processed strictly in the order you spoke them, and your text is never lost: if the AI is unavailable, the original text is inserted instead. The preview panel shows a live status ("text is being processed by AI…") so you always know what is happening. Recommended for long dictations. Without a key everything works as usual — the AI features simply stay inactive, and the settings clearly explain this. !! You pay your AI provider directly — no intermediaries, no extra service fees. Modern budget models (GPT-4o mini, Gemini flash, Claude Haiku, DeepSeek) cost fractions of a cent per correction, so even a $5 balance is enough to polish enormous amounts of text. !!

Your API Keys Are Safe

Keys are stored only on your computer, in encrypted form (AES-GCM), and are sent nowhere except your chosen AI provider. Web pages can never read your keys: the extension never exposes them to any site. The settings never display the key itself — only a masked hint (sk-…ab12). You can delete the key at any moment with one click, and nothing is saved until you press Save.

Functionality Disclaimer

Limitations of Speech Recognition APIs: Although speech recognition technologies have significantly improved, they still may struggle with accurately identifying punctuation. This is because punctuation marks are not spoken as words, and must be inferred from intonation, pauses, and the rhythm of speech, which can be challenging for algorithms. Complexity of Speech Models: Human speech can be unpredictable and varied. People may speak quickly, merge words, use regional features, or articulate sounds unclearly. All these factors pose challenges for the speech recognition system in accurately recognizing words and applying punctuation. Need for Contextual Understanding: Punctuation and word choice often depend on the context of a sentence or even a paragraph. AI models such as ChatGPT, Claude and DeepSeek are trained on large volumes of text and can better determine the right words, commas, periods, question marks, and exclamation points based on context. Brief Explanation for Inclusion in Texts: "To enhance the accuracy of texts recognized by Speech-to-Text technologies, we apply additional correction using AI. This allows for consideration of the context and nuances of human speech, ensuring high precision in text formatting."

Notes:

AI correction does not work for the YouTube comments section. On-the-fly improvement adds a short delay (1–3 seconds per phrase) — that is the price of accuracy; the status message on the panel keeps you informed.

Dictation in a Separate Window

For more convenient text management, use the dictate in a separate window feature:

Press the "Dictate window" button. Dictate your text, and it will appear in a separate field, from where you can copy it anywhere or use it for personal notes. Press "✨ Improve current text" to polish the whole text with your chosen AI provider right in this window (works great together with on-the-fly improvement, or on its own). Example of using: dictate an email to anybody and correct the text using AI.

Additional Features

All windows and panels of the plugin are draggable.

Hotkeys: To toggle the panel visibility, use the hotkey: CTRL+SHIFT+U. When using any textarea or text input, you can enable or disable speech recognition with the hotkey: CTRL+SHIFT+Y. If you select text and then speak, the selected text will be replaced with what you have spoken. Settings changes apply instantly in all open tabs. The interface is translated into 17 languages. Additionally, you can close the window or change the language settings at any time.

Feedback and Suggestions

If you have any ideas or suggestions, please contact us through contact form on https://pluginus.net/contact-us. We value your input and are always looking to improve "Voice Master" to better suit your needs.

Conclusion

"Voice Master" is a handy tool that helps you translate voice to text. With it, you can easily control the Internet with your voice, turn speech into text on command, and, if necessary, correct the text using artificial intelligence — instantly, as you speak, or afterwards with one click. This tool makes using the Internet easier and does not require much effort to manage. Add "Voice Master" to your daily routine to make texting easier with advanced speech recognition technology.