Voice Control
RemoteConsoleSSH lets you speak terminal commands instead of typing them. The app uses on-device speech recognition to capture your voice, then an AI model interprets your natural speech into the correct terminal command. This is especially useful when you need to run commands quickly without wrestling with the mobile keyboard.
Enabling Voice Input
Tap the microphone icon
The microphone button is located in the keyboard toolbar, above the on-screen keyboard. Tap it to activate voice input mode.
Grant microphone permission
On first use, your device will ask for permission to access the microphone. Tap Allow to continue. Speech recognition happens on-device -- your audio is not sent to a remote server.
Choose your mode
Select either Command mode or Text mode (explained below), then start speaking.
Info
Voice control is available on iOS and Android only. The web version does not support voice input due to browser limitations with speech recognition APIs.
Two Modes
Voice input operates in two distinct modes depending on what you are trying to do.
Command Mode (Default)
In Command mode, the AI interprets your natural speech as a terminal command. You speak in plain language and the app figures out the right command to run.
| What you say | What gets executed |
|---|---|
| "go to downloads" | cd ~/Downloads |
| "go to home" | cd ~ |
| "list all files" | ls -la |
| "show git status" | git status |
| "search for TODO" | grep -r "TODO" . |
| "show disk usage" | df -h |
| "find large files" | find . -size +100M |
| "check running processes" | ps aux |
The AI understands context and synonyms. "show me what's in this folder" and "list files" both produce ls -la. If the AI is unsure, it picks the most common interpretation and shows you the command before executing it -- you always get a chance to confirm or cancel.
Text Mode
In Text mode, your speech is transcribed as-is, with no AI interpretation. The raw text is sent to the terminal exactly as spoken. Use this mode when you are:
- Typing into a text editor (vim, nano)
- Entering input for an interactive prompt
- Working in a REPL (Python, Node.js, IRB)
- Composing a commit message
Tip
Switch between Command and Text mode with a single tap on the mode toggle in the voice input overlay. Command mode is better for running shell commands. Text mode is better for entering free-form text.
Full Voice Command Reference
Common Commands
| Voice Input | Terminal Command |
|---|---|
| "go to downloads" | cd ~/Downloads |
| "go to home" | cd ~ |
| "go to desktop" | cd ~/Desktop |
| "go back" | cd .. |
| "list files" | ls -la |
| "where am I" | pwd |
| "clear" / "clear screen" | clear |
| "git status" | git status |
| "git log" | git log --oneline |
| "git pull" | git pull |
| "git push" | git push |
| "run npm install" | npm install |
| "run npm start" | npm start |
| "make directory test" | mkdir test |
| "remove file old.txt" | rm old.txt |
| "show file contents" | cat <filename> |
| "edit file" | nano <filename> |
Special Keys by Voice
You can also trigger special key presses by voice. These work in both Command and Text mode.
| Voice Input | Action |
|---|---|
| "enter" / "return" / "submit" | Enter key |
| "control c" / "cancel" / "stop" | Ctrl+C (interrupt) |
| "tab" / "autocomplete" | Tab key |
| "escape" | Escape key |
| "up" / "previous command" | Up arrow |
| "down" / "next command" | Down arrow |
| "backspace" / "delete" | Backspace |
| "control d" | Ctrl+D (EOF / exit) |
| "control z" | Ctrl+Z (suspend) |
| "control l" | Ctrl+L (clear screen) |
Tip
Saying "control c" is a fast way to interrupt a runaway process when you do not want to reach for the keyboard toolbar. It works in both voice modes.
AI Provider Setup
The AI that interprets your voice commands in Command mode needs an API provider. You can choose from several options.
Open Settings
Go to Settings from the app's main navigation.
Navigate to Voice settings
Tap Voice and then AI Provider.
Choose a provider
| Provider | Notes |
|---|---|
| Google Gemini | Default option. Free tier available. Good general-purpose interpretation. |
| Anthropic Claude | Excellent at understanding ambiguous commands. Requires API key. |
| OpenAI | Widely used. Requires API key. |
Enter your API key
Paste your API key for the chosen provider. The key is stored securely in your device's keychain, just like your SSH credentials.
Info
The AI provider is only used for Command mode interpretation -- turning your natural speech into terminal commands. Speech recognition itself happens on-device and does not require an API key or internet connection for the transcription step.
The Voice Input Flow
Here is what happens when you use voice input in Command mode:
- You tap the microphone and speak: "show me all docker containers"
- On-device speech recognition transcribes your speech to text
- The text is sent to your configured AI provider for interpretation
- The AI returns the terminal command:
docker ps -a - The app shows you the interpreted command with a Confirm / Cancel prompt
- You tap Confirm and the command is executed in the terminal
You always see the interpreted command before it runs. Nothing is executed without your confirmation.
Platform Permissions
| Platform | Permission Required | When Prompted |
|---|---|---|
| iOS | Microphone access (NSMicrophoneUsageDescription) and Speech Recognition (NSSpeechRecognitionUsageDescription) | On first tap of the microphone button |
| Android | RECORD_AUDIO permission | On first tap of the microphone button |
| Web | Not supported | N/A |
If you accidentally deny the permission, you can re-enable it in your device's Settings app under the RemoteConsoleSSH entry.
Tips for Best Results
- Speak clearly and naturally -- You do not need to use exact command syntax. Say "go to the downloads folder" instead of "cd tilde slash downloads".
- Be specific when needed -- If the AI misinterprets a command, try being more specific. Instead of "delete it", say "remove file named old-backup.tar.gz".
- Use Text mode for input -- If you are typing into vim or a Python REPL, switch to Text mode so your speech is transcribed literally.
- Check the preview -- Always review the interpreted command before confirming. The AI is good but not perfect.
What's Next
- Keyboard toolbar -- Quick access to special keys and the voice button
- SSH connections -- Manage your server connections
- Persistent sessions with tmux -- Keep your sessions alive across disconnections