This repo lets you control your Windows® desktop by voice. Using keyboard, mouse, and also voice commands is intuitive, increases your productivity, and is really awesome.
Required is Windows 10 or 11, a headset, and the free Serenade voice recognition application.
- Download and install the free Serenade app from https://serenade.ai. During installation choose no plugin and no programming language, then go to Settings > Server and choose the 'Local' endpoint (good for privacy and speed). Close the application when finished.
- Download and unzip this repo from https://github.com/fleschutz/talk2windows/archive/main.zip.
- Open 'Windows PowerShell (Administrator)' and execute:
Set-ExecutionPolicy RemoteSignedto allow the execution of PowerShell scripts. - Open 'Windows PowerShell' and execute:
./setup.ps1in the unzipped talk2windows folder. It exports all voice commands with your personal wake word to Serenade. In the following we will use the wake word Windows.
- Put your headset on and check audio and microphone.
- Launch Serenade and click the slider to switch from Paused to Listening mode. Raise the mic audio level in case a 'low voice' is shown.
- Say: "Windows, open calculator" with a calm voice into the mic. This voice command executes the PowerShell script
scripts/open-calculator.ps1which launches the calculator app.
Most voice commands follow the generic scheme "<wake word>, <verb> <subject>" such as: "Windows, open calculator". Supported are the following voice commands:
Say hello by replacing <greeting> with: good afternoon, good evening, good morning, good night, hello, hey, hi, howdy, morning, or nice to meet you.
Install a Microsoft Store application by replacing <app> with: 7-Zip, CrystalDiskInfo, CrystalDiskMark, Discord, Disney+, Edge, Firefox, Google Chrome, IrfanView, Mp3tag, Netflix, Opera Browser, Opera GX, One Calendar, Quick Share, Rufus, Spotify, Sticky Notes, TikTok, Twitter, Visual Studio Code, VLC, Windows Terminal, or WhatsApp. To uninstall it simply say: "Windows, uninstall <app>".
Launch an installed application by replacing <app> with: 3D-Viewer, 7-Zip, Calculator, Character Map, CrystalDiskInfo, CrystalDiskMark, Discord, Disney+, Edge, Firefox, Git Extensions, Google Chrome, Magnifier, Microsoft Paint, Microsoft Store, Mp3tag, Netflix, Notepad, OBS Studio, One Calendar, One Note, Outlook, Quick Share, Remote Desktop, Screen Clip, Spotify, Sticky Notes, System Information, Thunderbird, Visual Studio, VLC, voice recorder, Windows Terminal, or WhatsApp. To close it simply say: "Windows, close <app>".
Launch the default Web browser (or a new tab) with the given website. Replace <name> with: Amazon, Apple, Baidu, BBC, Bing, BitBucket, BitBucket status, CDC, CIA, CNN, Discord, DistroWatch, Docker, Docker status, Dropbox, eBay, Facebook, FBI, FlightRadar, Flipboard, FourSquare, FRITZ!Box, FRITZ!Repeater, GitHub, GliderTracker, HolidayCheck, HRworks, Instagram, IPFS, Microsoft, NASA, NBC, Netflix, Notepad, Outdoor Active, PayPal, Pinterest, Pixabay, Plex, Serenade, Slashdot, Snap Store, Space Weather, SpaceX, Spotify, Starbucks, Tesla, TikTok, Toggl, Topo Map, Twitter, UFA, Unsplash, Walmart, WhatsApp, White House, Windy, Wikipedia, Wolfram Alpha, World News, Yahoo, YouTube, or Zoom. When finished say: "Windows, close tab".
Let the computer check something by replacing <item> with: audio, Bitcoin rate, CPU temperature, dawn, day, DNS, drives, dusk, Earth (fun), Ether rate, for crash dumps, my downloads, for rain, for snow, Git version, headlines, internet connection, internet speed, moon phase, month, New Year, operating system, Outlook, ping latency, recycle bin, Santa, swap space, temperature, Tether rate, time zone, up-time, VPN, weather, week, the wind, year, or zenith.
Press a keyboard shortcut to navigate in a browser window. Replace <shortcut> with: close tab, last tab, next page, next tab, previous page, previous tab, open new tab, open new window, reload page, scroll down, scroll to bottom, scroll to top, scroll up, tab 1, tab 2, tab 3, tab 4, tab 5, tab 6, zoom in, or zoom out.
Insert the given character at the current text cursor position. Just replace <char> with: at, backslash, comma, copyright, Dollar, Euro, hashtag, minus, number, percent, pipe, plus, Pound, slash, trademark, underscore, or Yen.
Insert text at the current text cursor position. Replace <text> with: bomb, bye bye, call?, checkmark, clapping, date, good-bye, good morning, heart, OK, shrugging, smiley, or thumbs up.
Launch the default Web browser (or a new tab) showing the given Google® service. Replace <service> with: Alerts, Blogger, Books, Calendar, Charts, Contacts, Docs, Drawings, Earth, Flights, Gemini, Keep, Mail, Maps, Mars, News, Notes, Photos, Play, Scholar, Search, Sheets, Shopping list, Sky, Slides, Stadia, Tasks, or Translate. When finished say: "Windows, close tab".
Launches the default Web browser (or a new tab) showing the given magazine website. Replace <name> with: Cosmopolitan, Focus, Golf, GQ, Health, National Geographic, People, RollingStone, Scientific American, Spiegel, Sports Illustrated, SZ, The Oprah, Time, Vanity Fair, Vogue, or Wired. When finished say: "Windows, close tab".
Launches the default Web browser (or a new tab) with the given manual. Replace <name> with: Ant, Apple, Audacity, Azure, Bash, Blender, Chrome, Edge, Firefox, GCC, Git, Jenkins, MSBuild, OBS Studio, PowerShell, Serenade, SystemRescue, Toyota, Vim, Voice (yes, this page!), Volkswagen, or Windows. When finished say: "Windows, close tab".
Launches the default Web browser (or a new tab) showing the given city in Google Maps. Replace <name> with the name of the city. When finished say: "Windows, close tab".
Launches the default Web browser (or a new tab) showing the given exchange rate. Replace <name> with: Bitcoin, Cardano, Ether, Euro, NFT collections, Solana, Tether, or US dollar. When finished say: "Windows, close tab".
Launches the default Web browser (or a new tab) and plays a music genre. Replace <genre> with: 70s, 80s, 90s, Alternative Rock, Blues, Chillout, Christmas, Classical, Country, Dance, Folk, Hip Hop, House, Indie, Jazz, Metal, Pop,Reggae, Relax, RnB, Rock, or Workout. When finished say: "Windows, close tab".
Starts the playback of the given sound, just replace <name> with: alarm, bee, beep, Big Ben, car, cat, cow, dog, donkey, elephant, elk, frog, goat, gorilla, horse, lion, parrot, pig, rattlesnake, vulture, or wolf.
Launches the default Web browser (or a new tab) and lets you play a free browser game. Replace <game> with: 2048, Agar, Catan Universe, Chess, Cube, ISS Docking Simulator, Pacman, Santa Tracker, Slither, Tetris, TicTacToe, Tower, Vanis, War Brokers, or Wordle. When finished say: "Windows, close tab".
Replaces the desktop background by a random photo from Unsplash.com. Replace <category> with: Animal, Beach, Car, Christmas, City, Desert, Fantasy, Military, Movie, Music, Nature, Plane, Random, Space, Sports, or Winter.
Sets a reminder, just replace <time> with: at # AM, at # PM, at midnight, at noon, at sunrise, at sunset, at tea time, in # minutes, or in # hours.
Launches the File Explorer showing the given drive. Replace <letter> with: C:, D:, E:, F:, G:, H: or M:. When finished say: "Windows, close file explorer".
Launches the File Explorer with the given folder. Replace <name> with: apps, applications, autostart, crash dumps, desktop, documents, downloads, Dropbox, home, music, OneDrive, pictures, recycle bin, repos, SSH, Talk2Windows, temporary, videos, or Windows. When finished say: "Windows, close file explorer".
Open the system settings with the given category - just replace <category> with: activation, apps, background, backup, bluetooth, color, date, default apps, developer, display, ethernet, lockscreen, maps, personal, printer, proxy, recovery, speech, start, system, taskbar, themes, time, update, USB, VPN, Wifi, or Windows. When finished say: "Windows, close system settings".
Shows the given table, just replace <table> with: bluetooth devices, city weather, earthquakes, emojis, environment variables, headlines, installed apps, installed software, installed voices, network adapters, printers, running processes, services, system info, tasks, time zones, or user groups.
Performs the given action to the computer, replace <action> with: hibernate, lock, reboot, shut down, or suspend.
Asks Windows a question, just replace <question> with: are you a computer?, are you here?, are you listening?, are you ready?, are you there?, are you with me?, can you talk?, do you listen?, how are you?, how do you do?, how much daylight today?, how old are you?, what about ...?, what is [abbreviation], what's the date today?, what's the temperature?, what's the time?, what time is it?, when is Christmas?, when is easter?, when is midnight?, when is noon?, when is President's day?, when is sunrise?, when is sunset?, when is tea time?, where am I?, where are you?, where is the ISS, who is your father?, or you can talk?.
Lets Windows execute a command, replace <command> with: calculate [expression], connect VPN, count down from [number], empty recycle bin, enable god mode, I need help, locate my phone, minimize all windows, ping [host], repeat last reply, roll a dice, save screenshot, set volume to [number] percent, shut up, sorry, spell [text], tell me a joke, tell me a quote, thank you, thanks, translate [text] to [language], turn volume off, turn volume on, turn volume up, turn volume fully up, turn volume down, update repository, or wakeup [hostname].
Say good-bye to your computer by replacing <farewell> with: bye, bye-bye, ciao, good-bye, I'll be back, see you, or see you later.
Send your email feedback to: markus.fleschutz [at] gmail.com
This open source project is licensed under the CC0 license. All trademarks are the property of their respective owners.