Voice Typing with Openai-Whisper

State-of-the-art, offline voice typing/real-time voice translation in Linux tty (or WFL sesson on Windows.) with a tiny bash script.

No grapical OS required. Yet the voice keyboard does work with GUI X, wayland, whatever. Speak text everywhere, like a boss.

• Privacy-focused. Uses Whisper AI or Whisper.cpp for offline speech recognition.
• Organic software (free, off-grid, open-source, unrestrictive licenses).
• Promotes peace through expediting communication and understanding.
• Hands-free utilizing sox for rudimentary sound-level detection (VAD).
• Leverages ydotool to type text into any active window with a virtual keyboard.
• Low memory requirements. Resources may be freed between each spoken interaction.

Caveats

Windowws 11. Unlike Windows® 11's voice keyboard (voice typing), which sends voice data to Microsoft™ for processing, this project keeps it all under your roof. Let's make that very clear!. This is a completely-independent project. If you still want to use Windows's voice keyboard, press Windows-H to set that up. And we'll see you next time!

When voice_typing detects speech, it trims unwanted background noise, and then loads Whisper for each spoken paragraph, which causes a noticeable wait before text appears. It is intended for occasional use. And it is the most economical on resources.

For longer sessions, instead of loading and unloading Whisper multiple times, we have added voice_client. It connects to your Whisper.cpp server. The server keeps running in the background, on the same machine. And it can be configured to start and stop with the app. Or it can run continuously on a dedicated server somewhere across the network. Try it. Users might discover significant speedup. :)

For slightly more-accurate, continuous, networked dictation with more features, try the whisper_dictation AI assistant project. Features include a conversational chatbot, AI image generation, and voice-controlled program launchers leveraging the full power of Python.

End feature creep. This project is just a starting point, and will remain so. There is no end to what you might do from here.

Requirements

• Whisper AI or Whisper.cpp
• ffmpeg
• sox
• lame
• ydotool
• jq
• tmux or screen (optional)
• curl (for clients)

Install Dependencies

This assumes Whisper AI or Whisper.cpp and dependencies are installed and working. Most are available through the official software update app for each platform. Please examine voice_typing and voice_client scripts and see how easy they are to customize for any occasion. They are around 50 lines is all. Do not run untrusted code.

Fedora/Centos:

dnf -y install sox curl lame ydotool jq
sudo cp /usr/lib/systemd/system/ydotool.service /etc/systemd/system/
sudo chmod +x /etc/systemd/system/ydotool.service
sudo systemctl daemon-reload
sudo systemctl enable ydotool
sudo systemctl start ydotool
sudo chmod +s $(which ydotool)

You might need Rpmfusion-freeworld installed to get versions of lame and sox that write mp3 files. sudo dnf install \ https://download1.rpmfusion.org/free/fedora/rpmfusion-free-release-$(rpm -E %fedora).noarch.rpm

The ydotool package has instructions in /usr/share/doc/ydotool/README.md where they say the man page may not be up to date.

Debian-based systems:

sudo apt install sox curl lame ydotool openai-whisper libsox-fmt-mp3 jq

If ydotool is not available, or you need a later version, snwfdhmp commented:

git clone https://github.com/ReimuNotMoe/ydotool
cd ydotool
mkdir build
cd build
cmake -DSYSTEMD_SYSTEM_SERVICE=ON -DSYSTEMD_USER_SERVICE=OFF ..
make -j `nproc`
sudo ln -s $(pwd)/ydotool /usr/local/bin/ydotool
sudo ln -s $(pwd)/ydotoold /usr/local/bin/ydotoold
sudo cp ./ydotoold.service /etc/systemd/system/
sudo systemctl daemon-reload
sudo systemctl enable ydotoold.service # later changed to ydotool.service
sudo systemctl start ydotoold.service

For quick troubleshooting

sudo systemctl status ydotoold  # Should show "Active: active (running)"
journalctl -u ydotoold -b | tail -n 20

Setup

Edit .bashrc and add the line, export YDOTOOL_SOCKET=/tmp/.ydotool_socket

Install voice_typing, make sure ydotool works, and it's good to go.

git clone https://github.com/themanyone/voice_typing.git
sudo systemctl enable ydotool.service
sudo systemctl start ydotool.service
ydotool type hello! # test
cd voice_typing
./voice_typing

Speak and text appears. No other interaction is required.

Launch client in tray

If there is a server avavailable, edit voice_client to change the server location from localhost to wherever it resides on the network.

Launch voice_client in text terminal. Or if there is a window manager available, install Guake to launch in tray with a hotkey: guake -e voice_client

Server setup

The whisper-cpp executable is available in the repos. But power users may prefer to compile Whisper.cpp with GPU optimizations for best results. We compiled using cmake -B build -DGGML_CUDA=1 -DWHISPER_SDL2=ON -DWHISPER_FFMPEG=yes for about 4x speedup. If it complains about unsupported compiler, the best option is to search for rpms for a CUDA-compatible version of gcc, such as g++-14.

To minimize resources, launch server with ggml-tiny.en.bin. It uses just over 111 MiB VRAM on our budget laptop. (48MiB with ggml-tiny.en-q4_0.bin quantized to 4Bits, which is also usable with no graphics card).

./whisper-server -l en -m models/ggml-tiny.en.bin --port 7777 --convert

This works, for testing. But for even better results, try install some Voice Activity Detection (VAD) and start whisper-server that way, e.g.

whisper-server -l en -vm models/ggml-silero-v6.2.0.bin --vad -m models/ggml-base.en-q5_1.bin --convert --port 7777

Industrial client/server on same machine

For the professional at home, at work, and on the go, we have voice_client_local pre-configured for noisy environments, with a couple added lines to start and stop the server with the app:

systemctl --user start whisper.service

cleanup () {
    systemctl --user stop whisper.service

To avoid using sudo, configure whisper server to run as a user service:

cat <<EOF>$HOME/.config/systemd/user/whisper.service
Description=Run Whisper server
Documentation=https://github.com/openai/whisper

[Service]
ExecStart=$HOME/.local/bin/whisper-server \
-m $HOME/Downloads/src/whisper.cpp/models/ggml-medium-q5_0.bin \
-sns --convert --port 7777

[Install]
WantedBy=default.target
EOF

This will create $HOME/.config/systemd/user/whisper.service which should be edited for the preferred port, model, vad, etc. We're still using port 7777 to avoid conflicts.

Next, run systemctl --user daemon-reload to make the service available.

Install Guake and launch guake - e voice_client_local with a hotkey, such as Alt-Ctrl-V.

Other terminals

Instead of launching in the tray with Guake, the script may be launched with xterm e.g. xterm -e voice_client_local or any other terminal. It does not actually need a desktop to run. But if it does run on the desktop, we can use ydotool to press Alt-Tab. That should return control to whatever desktop app was running before it was launched. Add a line like this.

sleep 0.25 && ydotool key 56:0 42:0 56:1 15:1 56:0 15:0

Notes

• Adjust mic volume for best results. If recording never stops, mic volume is up too high. If you can't adjust volume for some reason, edit voice_typing or voice_client. And change silence-detection threshold from 4% and 2% to something higher. rec -c 1 -r 22050 -t mp3 "$tmp" silence 1 0.2 6% 1 1.0 5%

• Optionally create a Keybinding for mic mute/unmute. If there is continuous noise in the background, it goes into a recording loop and never gets around to typing text.

• First run of voice_typing might be slow as it needs to download the model (better yet, use whisper or whisper.cpp from cli first to download the model (tiny))

Troubleshooting

"failed to connect socket `/tmp/.ydotool_socket': Permission denied" Error

When encountering the error "failed to connect socket `/tmp/.ydotool_socket': Permission denied," it's essential to ensure that the current user has sufficient permissions to access the socket file. Here are some steps to troubleshoot this issue:

Check User Permissions and Service Status. Ensure that the user has been added to the "input" group and has the necessary permissions to access the socket file. Verify the status of the ydotool service to ensure it is running as expected.

Setuid Bit on the Executable. Consider setting the setuid bit on the ydotool executable using the command:

sudo chmod +s $(which ydotool)

This step can help address permission issues when running ydotool as a user.

Address Already in Use. If encountering the error "failed to bind socket: Address already in use," it may be necessary to delete the socket file from /tmp to resolve the issue.

Linking to the Expected Socket. If ydotool started as a user looks for the socket "/run/user/1000/.ydotool_socket" but the daemon as a systemwide service listens to /tmp/.ydotool_socket, consider creating a link to the expected socket to ensure proper functionality.

Report others issues in the GitHub issue tracker.

Similar Projects

• Whisper Typer Tool
• Whisper Dictation
• BabelFish Universal Translator

Thanks for trying out Voice Typing!

• GitHub https://github.com/themanyone
• YouTube https://www.youtube.com/themanyone
• Mastodon https://mastodon.social/@themanyone
• Linkedin https://www.linkedin.com/in/henry-kroll-iii-93860426/
• Buy me a coffee https://buymeacoffee.com/isreality
• TheNerdShow.com

Copyright (C) 2026 Henry Kroll III, www.thenerdshow.com. See LICENSE for details.