Give your AI the whole idea.
Say the context, the details, and the “one more thing.” Get your thoughts into a prompt without typing every word.
YOU, THINKING OUT LOUD
AI PROMPTS“Here’s what I’m trying to build…”
VOICE PRODUCTIVITY FOR WINDOWS
Speak your draft.
Hear it back.
Voice input for AI prompts
and everyday writing.
Dictate the context for an AI prompt, an email, or a note. Then listen to selected text read aloud. Speech processing stays on your Windows PC.
Windows 10 21H2+/11 x64 English 454.8 MB
Unsigned beta · Compatible Vulkan GPU required · Setup detailsA thought. A first draft.
Your next draft
FROM A THOUGHT TO A PROMPT
Help me turn this idea into a plan. I want to start a small business, and here’s what I’m thinking…
Your words, ready to use
Illustrative demo. No microphone access needed.
Speech stays on your PC
Dictate offline after setup
Dictation and read aloud
START WITH YOUR VOICE
From a thought
to a first draft.
The first draft is already in your head.
Talkasaurus helps you get it onto the screen.
Say what you meant to send.
Get an email or message started in your own words. Review the text, make it yours, and send when you’re ready.
THAT REPLY YOU’VE BEEN MEANING TO WRITE
EMAIL AND MESSAGES“Hi Alex, thanks for your patience…”
Catch it before it’s gone.
A paragraph, a reminder, or the beginning of something good. Capture the thought while it’s still fresh.
YOUR NEXT GOOD IDEA
NOTES AND IDEAS“A few things to remember for tomorrow…”
HOW DICTATION WORKS
A shortcut. Your voice. Your words.
Once you’ve set up your microphone and speech model,
your next thought is just a shortcut away.
- F9press & hold01
Hold your shortcut
Click into a text field and hold your talk key. Prefer a toggle? Start and stop with a tap instead.
- say what’s on your mind02
Let your thoughts out
Speak naturally. Talkasaurus captures your voice and turns it into text using a model on your PC.
- And just like that, words.release & review03
Keep your flow going
Release the key. When transcription finishes, your words appear in the active text field. Review and carry on.
Make it your own with configurable keyboard, mouse, or compatible foot-pedal shortcuts.
SEE TALKASAURUS IN ACTION
Speech to text.
Text to speech.
Dictate into the app you’re using, or listen as Talkasaurus reads selected text aloud.
01 SPEECH TO TEXT
Speak naturally. Keep your flow.
Tap your configured start/stop shortcut to dictate into the active text field. Tap it again to finish, and your words appear for review.
02 TEXT TO SPEECH
Select text. Listen at your pace.
Have Talkasaurus read a selected passage aloud. Press “Hear a voice sample” to listen to the same voice used in the app.
The animations use configurable example shortcuts; your saved shortcuts take priority. The speech-to-text animation is illustrative. No microphone access is needed on this website.
YOUR VOICE IS YOUR BUSINESS
Your voice.
Your Windows PC.
Talkasaurus processes your speech right on your computer. Your recordings don’t need to travel to a cloud transcription service.
- Dictation works offline after the initial setup.
- Talkasaurus doesn’t save recordings or a transcript history.
- You choose when to start and stop dictating.
Where your words go next is up to you. Text you put into an online app is handled according to that app’s privacy practices.
A FEW GOOD QUESTIONS
Before you
try the beta.
Here’s what to know
before you try the beta.
How do I try Talkasaurus?
Download the free Windows beta, run the installer, and follow the setup wizard to choose a speech model and microphone. The installer is 454.8 MB; speech models download separately. Start with Hatchling, approximately 148 MB, plus a small speech detector. Setup tests the model on your GPU before you begin.
This is an early beta. If you run into a problem, email [email protected] with your app version and a description of what happened.
Which computers does it support?
Designed for Windows PCs with compatible AMD, NVIDIA, or Intel graphics. Setup checks your hardware and helps you choose a transcription model.
This beta supports Windows 10 version 21H2 (build 19044) or newer and Windows 11 on x64 PCs, with English dictation. Version 0.7.33 uses Vulkan, a shared GPU interface supported by compatible AMD, NVIDIA, and Intel graphics cards. A working graphics driver with Vulkan support is required. Setup runs a speech sample before accepting a model. Hardware testing is still limited, and performance depends on your GPU and the model you choose. Windows ARM is not supported.
macOS and Linux may be considered in the future, but we’re focused on Windows first. A mobile app is not planned.
Does it need an internet connection?
You’ll normally download the app and a speech model during setup; a model can also be imported from a local file. Once set up, speech-to-text processing works on your PC without an internet connection. Downloads and updates require connectivity.
Does it work with my AI assistant or email app?
Talkasaurus inserts transcribed text into the active text field, so you can use it for prompts, messages, and documents in compatible Windows apps and browsers. Some protected or elevated fields may prevent insertion, and compatibility is still being tested during the beta.
It doesn’t send a message or submit a prompt for you. Review your words before sending.
Can I use a foot pedal?
Yes, a compatible pedal that sends a keyboard key can be assigned to your Talkasaurus shortcut. You can also use configurable keyboard and supported mouse-button shortcuts, with hold-to-talk or start/stop modes.
What will it cost?
The beta is free. Pricing for the commercial release is coming soon; we haven’t decided on a price or payment model yet.
THE FREE WINDOWS BETA IS HERE
Bring your next draft
to life with your voice.
Try local dictation and read-aloud playback on your Windows PC.
The beta is free while we continue testing and improving it.
Download free Windows betaVersion 0.7.33 454.8 MB
Commercial pricing coming soon
Commercial pricing coming soon
Designed for Windows PCs with compatible AMD, NVIDIA, or Intel graphics. Setup checks your hardware and helps you choose a transcription model.
Windows 10/11 x64 · English dictation.
Speech models download separately. Start with Hatchling.
This beta installer is not digitally signed. Windows may show an unknown-publisher warning.
Download SHA-256 checksum Verify before installing
A good place to start.
Product facts
Local dictation, read-aloud playback, requirements, and beta limits.
Set up your Windows PCInstall, choose a model, test your microphone and find help.
Try a workflowDraft an AI prompt, email or note, then review your words.
Reading on your phone? Open the setup guide on your Windows PC later. This page does not check your hardware.