Press a hotkey anywhere on your system, speak naturally, and Telvr transcribes your voice in real-time using Whisper. The finished text is automatically inserted at your cursor position — no copy-paste, no app switching.

Which languages are supported?

Telvr supports 50+ languages via OpenAI's Whisper large-v3 model. Language detection is automatic — just speak in your preferred language and Telvr handles the rest.

Do I need a subscription?

No. Telvr uses a pay-as-you-go model: EUR 3 per month infrastructure fee plus EUR 0.03 per minute of usage. No lock-in, no auto-renewal. You top up your balance and use it at your own pace.

Does Telvr work offline?

Currently, Telvr requires an internet connection for cloud-based transcription via Groq. A Community Edition with local processing using your own API key is planned for the future.

Which apps are supported?

Telvr works system-wide — it inserts text at the cursor position in any application. Email clients, chat apps, code editors, browsers, word processors — if you can type in it, Telvr works there.

All data is transmitted via TLS encryption. Audio recordings are not permanently stored after transcription. Groq processes your audio under a Data Processing Agreement (DPA). We do not sell or share your data.

← Blog2026-02-17

Speech Recognition on Windows: Complete Setup Guide (2026)

Windows Voice Input Options

Windows offers more speech recognition options than any other desktop platform, ranging from the completely free built-in tools to enterprise-grade professional software. Setting up the right option takes anywhere from two minutes to an afternoon depending on what you need.

This guide covers three tiers: the built-in Windows Voice Typing that is ready immediately, Dragon Professional for specialized professional use, and the current state of modern AI-powered options for Windows users who want more than the built-in tool.

Option 1: Windows Voice Typing (Built-in)

Windows Voice Typing is the fastest path to voice input on Windows. It requires zero installation and is available on Windows 10 version 20H2 and later, as well as Windows 11.

Setup

Press Win + H from any application with an active text field
A microphone widget appears at the top of the screen
Click the microphone button or press Win+H again
Start speaking

That is it. No installation, no account, no configuration required for basic use.

Enabling Auto-Punctuation

Auto-punctuation (Windows 11 and recent Windows 10 builds) adds commas and periods automatically without you having to say them explicitly.

Press Win+H to open Voice Typing
Click the gear icon in the widget
Enable "Auto-punctuation"

Language Setup

If you want to dictate in a language other than your system default:

Go to Settings > Time and Language > Language and Region
Add your desired language pack
Download the speech recognition package for that language
Switch input language via the taskbar language selector before dictating

Supported languages for Windows Voice Typing (as of 2026): English (US, UK, AU, IN), Chinese (Simplified, Traditional), French, German, Spanish, Italian, Portuguese, Japanese, Korean, Arabic, and approximately 15 others. Full list at Microsoft's official documentation.

Voice Typing Commands

Windows Voice Typing supports several voice commands:

"Stop listening" — deactivates the microphone
"Delete that" — removes the last dictated text
"Scratch that" — removes the last utterance
"Go to sleep" / "Wake up" — toggles listening state

Limitations

Windows Voice Typing produces raw transcription. There is no AI enrichment, no email formatting, and no structured output. Filler words appear in the text. Punctuation is handled by auto-punctuation when enabled, but complex sentence structures may need manual cleanup.

It works well for everyday dictation where you expect to edit the output. It is not the right tool if you need professional-quality text without editing.

Option 2: Dragon Professional (Windows)

Dragon Professional is the choice for Windows users who need professional-grade accuracy with domain-specific vocabulary — primarily in law, medicine, finance, and technical fields.

Installation

Purchase Dragon Professional from Nuance's website ($699 as of 2026)
Download the installer
Run the setup wizard
Create a voice profile (initial training takes 5-10 minutes)

Voice Profile Training

Dragon's voice training improves accuracy significantly. During setup:

Read provided text passages (2-3 minutes of reading)
Dragon analyzes your voice patterns, accent, and speaking style
The profile saves to your user account

Accuracy improves further over time as Dragon learns corrections you make.

Custom Vocabulary

The custom vocabulary feature is Dragon's strongest selling point for professional use:

Open Dragon's vocabulary editor
Add domain-specific terms (medical conditions, legal citations, product names, technical terms)
Add written form and spoken form for abbreviations (spoken: "the company," written: "Acme Corporation")

For professionals with highly specialized vocabulary, this feature alone justifies the cost.

Dragon Basics

The push-to-talk equivalent in Dragon: press and release a customizable key (default is the minus key on the numpad) to start listening. Say "stop listening" or press the key again to stop.

Dragon also supports application control commands — you can dictate to switch between applications, click buttons, navigate menus, and control Windows features by voice. This is useful for users with mobility impairments.

Option 3: Modern AI Speech Recognition for Windows

The gap in the Windows market in 2026 is the absence of a tool that combines Whisper-level transcription accuracy with AI text enrichment in a system-wide push-to-talk interface. Tools with that combination exist on macOS (Telvr, Wispr Flow) but not yet on Windows.

What Windows Users Can Use Today

Whisper Desktop (open-source): Several community tools bring Whisper to Windows. The most maintained are:

Whispering (open-source, GitHub) — records audio, transcribes via local Whisper, pastes to clipboard
FasterWhisper on Windows — requires Python setup, better performance via quantized models

These produce raw Whisper transcription without enrichment. Setup requires technical comfort with command-line tools.

Voice In (Chrome Extension): A browser extension that adds voice input to any text field in Chrome. Uses Google's Web Speech API, not Whisper. Accuracy is good for English; limited for other languages. Works only inside Chrome.

Telvr for Windows

Telvr's Windows version is in development. When available, it will bring the full Telvr experience to Windows: push-to-talk, Whisper large-v3 transcription, six AI enrichment modes, and system-wide text insertion.

Microphone Setup (Applies to All Methods)

Microphone quality matters more than which speech recognition tool you use. A poor microphone will limit accuracy regardless of the underlying model.

Built-in Laptop Microphone

Workable in quiet environments. May struggle with background noise, HVAC, or ambient office sounds.

External USB Microphone

A significant upgrade. Entry-level USB mics ($50-100, Blue Yeti Nano, Rode NT-USB Mini) produce substantially cleaner audio than built-in laptop mics. The improved input quality translates directly to better accuracy.

Headset Microphone

Consistent microphone distance and position makes headsets particularly good for dictation. USB or Bluetooth headsets both work; avoid analog 3.5mm if possible as they tend to introduce noise.

Microphone Setup in Windows

Go to Settings > System > Sound
Under Input, select your preferred microphone
Click "Configure microphone" and follow the calibration wizard
Use the volume meter to verify the input level is strong without clipping

Target input level: The volume bar should reach roughly 75% of maximum when speaking at a normal conversational volume. Too quiet means weaker recognition; too loud causes clipping.

Troubleshooting Common Issues

Voice Typing not activating (Win+H):

Check that a text field is active (click into a text box first)
Verify microphone permissions: Settings > Privacy and Security > Microphone

Poor accuracy in any tool:

Test with a different microphone
Move to a quieter environment
Speak slightly more slowly and clearly
For Dragon: run the accuracy tuning wizard after several hours of use

Text appearing in wrong locations:

The target application may not support voice input in that specific field
Ensure focus is on the correct window and text field

High latency (long wait between speaking and text appearing):

Windows Voice Typing processes on Microsoft's servers; check internet connection
For Dragon: local processing, check CPU usage — other demanding apps can slow it down