CamoVoice User Guide

Fully Private, Offline Speech-to-Text

Looking for CamoVoice International?

CamoVoice International supports transcription in English, Español, Deutsch, Français, Italiano, Português, Nederlands, Polski, Svenska, and Русский.

Jump to International Guide

Overview

CamoVoice is a desktop speech-to-text application designed with one core principle: your voice stays on your device. Unlike cloud-based transcription services, CamoVoice processes everything locally using bundled AI models — no internet connection required, no data leaves your computer, and absolutely zero telemetry.

Privacy & Architecture

100% Offline: All transcription happens on your machine using faster-whisper, which runs entirely locally
Zero Telemetry: CamoVoice collects nothing. No usage analytics, no crash reports, no audio samples
No Cloud Dependencies: Once installed, the app works without any network connection
Local Settings: Your preferences are stored in a simple settings.json file on your device and auto-loaded at open
English Optimized: Purpose-built for English transcription with maximum accuracy

This architecture makes CamoVoice ideal for transcribing sensitive notes, legal dictation, personal journals, or any audio where privacy matters.

User Interface

CamoVoice features a high-contrast, dark-themed interface designed for simplicity and accessibility.

Main Window Layout

┌─────────────────────────────────────────────────────────────┐
│  [Load Audio File]          🎙 Record Button     Status     │
│  [Settings]        "Click or hold spacebar"     🎤 Device   │
├─────────────────────────────────────────────────────────────┤
│                                                             │
│                   Transcription Output                      │
│                      (editable text)                        │
│                                                             │
├─────────────────────────────────────────────────────────────┤
│  [Copy Text] [▶ Play] [Save As]       [Undo Clear] [Clear] │
├─────────────────────────────────────────────────────────────┤
│  [Input Audio: ████████░░] ← (appears during recording)     │
├─────────────────────────────────────────────────────────────┤
│  Text Size: ═══════●═══════    Mode: [Fast|Thinking]        │
└─────────────────────────────────────────────────────────────┘

 ● REC  ▌▌▌▌▌░░   ← live status pill (top-left of primary screen
                    while the global voice-typing hotkey is active)

Recording Audio

There are two ways to record:

Click the Microphone Button: Click once to start, click again to stop
Hold Spacebar: Hold spacebar to record, release to stop (push-to-talk style)
- A brief 200ms delay prevents accidental recordings from quick taps
- If you're typing in the text area, spacebar works normally

During recording, the microphone button pulses and fills with the accent color. When you stop, transcription begins automatically.

Global Voice Typing Hotkey New

CamoVoice can also record and paste transcriptions into any other application — no need to have the CamoVoice window focused. In Settings, enable “Enable global hotkey (Ctrl+` / ⌃`)”, then:

Place your cursor in any text field (email, chat, document, browser, IDE, etc.).
Press Ctrl + ` on Windows or ⌃ + ` on macOS to start recording.
A small dark REC pill appears in the top-left of your primary display with a blinking indicator and a live audio-level bar, so you always know the mic is hot.
Press the same shortcut again to stop. CamoVoice transcribes locally, copies the result to your clipboard, and automatically pastes it at your cursor.

The global hotkey is disabled by default and only activates while CamoVoice is running. It is intentionally set to Ctrl+` (backtick) rather than Ctrl+Space or Ctrl+Shift+Space so it doesn't collide with macOS Spotlight, the input-source switcher, or common editor shortcuts.

On macOS, global voice typing also requires one-time Accessibility and Input Monitoring permission (steps in Settings → macOS Permissions for Global Hotkey below).

Privacy note: Transcription still happens entirely on your device. Nothing about the global hotkey changes CamoVoice's offline-first architecture — there is no background listening when the hotkey is idle, and no audio is ever transmitted.

Recording Duration Timer

For longer recordings, a duration timer appears after 1 minute of recording, showing your elapsed time and the mode's limit (e.g., 3:45 / 10:00). The timer changes color as you approach the limit:

Gray: Normal — plenty of time remaining
Orange: Warning — less than 1 minute remaining
Red: Critical — less than 30 seconds remaining

This helps you pace your recordings and avoid hitting the automatic cutoff unexpectedly.

Transcription Animation

When CamoVoice is processing your audio, you'll see an animated "Transcribing..." status with cycling dots, and the text area border will pulse with the accent color. This provides clear visual feedback that your audio is being processed.

Audio Level Indicator

While recording, CamoVoice displays a thin Input Audio meter below the action buttons that shows real-time audio levels:

Orange gradient bar: Shows current audio input level
Yellow peak indicator: Marks the highest audio peaks
Automatic display: Appears when recording starts
Smart hiding: Disappears 1.5 seconds after recording stops

This visual feedback confirms your microphone is working and picking up audio.

Loading Audio Files

Click Load Audio File or drag and drop a supported file onto the CamoVoice window to transcribe existing audio. CamoVoice supports:

WAV — Any sample rate or bit depth
MP3 — Decoded natively with the bundled miniaudio library (no ffmpeg required)
M4A / AAC — Decoded via bundled ffmpeg in packaged builds

The app automatically resamples to the required 16kHz and converts stereo to mono. Files currently have a 90 MB (Fast mode) and 45 MB (Thinking mode) size limit to protect responsiveness on typical consumer laptops.

Note: M4A/AAC support in packaged builds is provided by a bundled ffmpeg binary and works fully offline after installation. MP3 support is fully offline without any external binary.

Transcription Output

The large text area displays your transcriptions. Each new recording or file appends to existing text, making it easy to build up longer documents.

Working with Text

Edit freely: Click into the text area to make corrections or additions
Copy Text: Copies all text to your clipboard (button flashes green to confirm)
▶ Play: Reads the transcription aloud using text-to-speech (see Playback Voice below)
Save As: Opens a save dialog with timestamp options (see below)
Undo Clear: Restores the most recently cleared text (grayed out when nothing to undo)
Clear: Removes all text and timestamps

Undo Clear

If you accidentally clear your text, you have two options:

Click the Undo Clear button (located to the left of Clear)
Press Ctrl+Z (Windows) or ⌘+Z (macOS)

This restores both the text and any associated timestamps from the most recent clear action. The Undo Clear button is grayed out when there's nothing to restore.

If the undo itself was a mistake and you want to re-clear the restored text, press Ctrl+Shift+Z (Windows) or ⇧+⌘+Z (macOS). The text and timestamps remain available for undo afterward, so you can freely toggle between the two states.

Save As Options

When you click Save As (or press Ctrl+S / ⌘S), a dialog appears with options:

Include date at top of document: Adds a header with the current date
Include timestamp after each transcription: Adds a [Recorded: DD-Mon-YYYY HH:MM:SS] line after each transcription segment (e.g., [Recorded: 09-Jan-2026 14:32:15])
Include edits to transcription: Saves the text exactly as it appears in the app, including any spelling corrections or edits you've made. When this is enabled, timestamps are disabled since edited text may no longer match the original transcription segments.

Export Formats

CamoVoice supports three export formats:

Format	Extension	Best For
Text File	`.txt`	Universal compatibility, email, plain text editors
Word Document	`.docx`	Professional documents, further editing in Microsoft Word
PDF Document	`.pdf`	Sharing, archiving, printing (read-only format)

Simply select your desired format from the "Save as type" dropdown in the file dialog.

Example output with both timestamp options enabled:

Transcription Record - 09-Jan-2026
========================================

This is my first transcription segment.
[Recorded: 09-Jan-2026 14:32:15]

This is my second transcription segment.
[Recorded: 09-Jan-2026 14:35:42]

Transcription Modes

CamoVoice offers two transcription modes, selectable at the bottom of the window:

Model Mode	Speed	Accuracy	Best For
Fast	★★★★★	★★★☆☆	Quick notes, short recordings, real-time feel
Thinking	★★★☆☆	★★★★★	Important transcriptions, longer recordings

Fast mode: Uses the compact base model (~140MB) with greedy decoding for maximum speed. Great for quick voice memos, short dictation, or when you want near-instant results. Trades some accuracy for responsiveness.
Thinking mode: Uses the full large model (~1.4GB) with beam search and voice activity detection. Better accuracy for complex vocabulary, accents, background noise, or professional transcription work.

Your mode selection is automatically saved. The model loads on first use of each mode, so the first transcription after opening the app or switching modes may take slightly longer.

Recordings have a time limit (currently 15 minutes for Fast, 7 minutes for Thinking) to protect responsiveness; recording auto-stops and transcribes at the limit.

Settings

Click the Settings button to open the settings panel.

macOS Permissions for Global Hotkey

After downloading the DMG and installing CamoVoice, the first time you record you will be prompted to allow microphone access. Click Allow. In-app recording should then work immediately using your default microphone (or the input device you selected in Settings).

To enable global hotkey recording / voice typing in text fields outside the app (⌃+` or Ctrl+` depending on keyboard), complete these one-time macOS steps:

Open System Settings → Privacy & Security → Accessibility.
Click +, select CamoVoice from Applications, and ensure it is toggled on.
Open System Settings → Privacy & Security → Input Monitoring.
Click +, select CamoVoice from Applications, and ensure it is toggled on.
Quit and re-open CamoVoice.
Test hotkey recording in a text window outside the app.

Input Device

Select which microphone to use for recording. Options include:

Default: Uses your system's default input device
All detected input devices are listed with their audio API

If you have multiple microphones (e.g., headset, webcam, USB mic), select the one you want to use here.

Playback Voice

Choose the voice used for the ▶ Play feature, which reads your transcription aloud.

These are system voices installed on your device, not voices bundled with CamoVoice. To add or remove voices:

Windows: Settings → Time & Language → Speech → Manage voices
macOS: System Settings → Accessibility → Spoken Content → System Voice → Manage Voices

The available voices depend on what you've installed on your operating system.

Show Timestamps in Transcriptions

When enabled, each transcription segment displays its recording time directly in the app, formatted like a 24-hour ISO-style legal/medical record with unambiguous month abbreviations:

Your transcribed text here.
[Recorded: 09-Jan-2026 14:32:15]

For live in-app recordings, the timestamp reflects when the recording was made. When you load or drag and drop an audio file for transcription, CamoVoice uses the timestamp embedded in the audio file (when available) instead of the time of transcription.

This is useful for keeping track of when recordings were made. Timestamps are always stored internally — this setting just controls whether they're visible in the app. When you save, you can choose whether to include timestamps regardless of this display setting.

Keep Window Always on Top

When enabled, the CamoVoice window stays visible above other applications. This is useful for:

Dictating while reading documents or browsing
Keeping CamoVoice accessible during multitasking
Hands-free workflows where you can't easily switch windows

Toggle this in Settings → "Keep window always on top".

Custom Words

CamoVoice allows you to add custom vocabulary — specialized terms, names, phrases, jargon, or unusual words that you frequently use. This helps the speech recognition model recognize these words more accurately.

To manage custom words:

Open Settings
Click Custom Words (n) (where n is your current word count)
In the Custom Words window:
- Type a word or phrase and click Add (or press Enter)
- View all your current custom words in the scrollable list
- Click the ✕ next to any word to remove it
- Click Clear All to remove all custom words
- Click Done when finished

Important notes:

Custom words are suggestions that increase recognition likelihood — they are not guaranteed to be transcribed exactly
You can add up to 100 custom words
Words are case-preserved but duplicate-checked case-insensitively
Custom words are saved automatically and persist between sessions

Best uses for custom words:

Proper names (people, companies, products)
Technical jargon or industry-specific terms
Acronyms that should be recognized as spoken
Unusual spellings or foreign words used in English context

Accessibility Features

Scalable Interface

The Text Size slider adjusts the font size throughout the entire application — not just the transcription area. This includes:

Transcription text
All buttons and labels
Status messages
Mode selector

Slide right for larger text, left for smaller. Your preference is automatically saved.

This makes CamoVoice usable for:

Users with visual impairments
High-DPI displays
Presentations or screen sharing
Personal comfort preferences

Keyboard Shortcuts

Action	Windows	macOS
Hold to record (in-app)	`Spacebar` (hold)	`Spacebar` (hold)
Global voice-typing toggle (system-wide)	`Ctrl`+`	`⌃`+`
Save As	`Ctrl`+`S`	`⌘`+`S`
Undo clear	`Ctrl`+`Z`	`⌘`+`Z`
Redo clear (revert undo)	`Ctrl`+`Shift`+`Z`	`⇧`+`⌘`+`Z`
Close dialogs	`Escape`	`Escape`

Note: The Escape key closes Settings and Save As dialogs without saving changes.

How It Works (Technical)

CamoVoice is built on several key technologies:

Speech Recognition

Two model options: Fast mode uses base.en (~140MB), Thinking mode uses distil-large-v3 (~1.4GB)
faster-whisper: CTranslate2-based implementation that's 4-6x faster than the original Whisper with lower memory usage
Models are bundled with the application — no download required after installation
All processing uses your CPU with int8 quantization for efficiency

Audio Capture

SoundDevice: Cross-platform audio recording
Records at 16kHz mono (Whisper's native sample rate)
Audio is processed directly in memory — no temporary files

Text-to-Speech

pyttsx3: Interfaces with your operating system's speech synthesis
Uses voices installed on your system (SAPI5 on Windows, NSSpeechSynthesizer on macOS)

User Interface

CustomTkinter: Modern-looking Python GUI framework
Dark theme with orange accent colors
Responsive layout that scales with window size (minimum 640×480)

Troubleshooting

"No voices available" for Playback

Your system may not have TTS voices installed. Install voices through your operating system's speech settings.

Recording doesn't start

Check that your microphone is connected and selected in Settings
Ensure no other application is exclusively using the microphone
Try selecting a different Input Device

Transcription is inaccurate

Speak clearly and at a moderate pace
Reduce background noise
Try switching to Thinking mode for better accuracy
Ensure you're speaking English (CamoVoice is optimized for English)

App is slow on first transcription

The first transcription may take slightly longer as the model initializes. Subsequent transcriptions will be faster. Switching modes also requires loading a different model.

Tips for Best Results

Speak clearly with natural pacing — the model handles conversational speech well
Minimize background noise — even with noise handling, clearer audio = better results
Use Fast mode for quick notes or short sequential dictations, Thinking mode for important uninterrupted long transcriptions
Edit as you go — freely adjust the text manually to fix mistakes or add/remove text before or after transcriptions

Enterprise Customization & Expansion Options

17 input languages: English, Spanish, French, German, Italian, Portuguese, Dutch, Russian, Chinese, Japanese, Korean, Polish, Turkish, Ukrainian, Arabic, Hindi, and Vietnamese
Option to auto-detect input language
Option to auto-translate to English
Additional file formats for input sounds and output files
Enhanced audit trail features and export formats

Privacy Commitment

✓ No internet connection required after installation
✓ No accounts, no sign-ups, no authentication
✓ No analytics or usage tracking
✓ No audio sent to any server
✓ Settings stored locally in plain JSON
✓ Models bundled locally — no APIs, no third party exposure

CamoVoice v1.0.0

CamoVoice International User Guide

Fully Private, Offline Multilingual Speech-to-Text

Overview

CamoVoice International is a desktop speech-to-text application designed with one core principle: your voice stays on your device. Unlike cloud-based transcription services, CamoVoice processes everything locally using a bundled multilingual AI model — no internet connection required, no data leaves your computer, and absolutely zero telemetry.

Privacy & Architecture

100% Offline: All transcription happens on your machine using faster-whisper, which runs entirely locally
Zero Telemetry: CamoVoice collects nothing. No usage analytics, no crash reports, no audio samples
No Cloud Dependencies: Once installed, the app works without any network connection
Local Settings: Your preferences are stored in a simple settings.json file on your device and auto-loaded at open
Multilingual: Transcribes English, Spanish, German, French, Italian, Portuguese, Dutch, Polish, Swedish, and Russian — or auto-detects the language from your audio

This architecture makes CamoVoice ideal for transcribing sensitive notes, legal dictation, personal journals, or any audio where privacy matters.

Multilingual Support

CamoVoice International supports transcription in ten languages out of the box, plus automatic language detection from the audio itself.

Supported Languages

Language	Code	Interface Language Available
Auto-detect	—	— (UI stays in English)
English	en	✓
Español	es	✓
Deutsch	de	✓
Français	fr	✓
Italiano	it	—
Português	pt	—
Nederlands	nl	—
Polski	pl	—
Svenska	sv	—
Русский	ru	—

Choosing a Language

Open Settings and use the Input Language dropdown. The selection does two things:

Transcription: The selected language is passed to faster-whisper so it knows which phonemes to expect. This generally improves accuracy over auto-detect when you know the language up front.
Interface language: If you choose English, Spanish, German, or French, the main-window buttons, Settings window, Custom Words window, Save As dialog, and common status messages re-label into that language. Other languages transcribe correctly but keep the interface in English.

Auto-Detect

If you leave Input Language on Auto-detect, faster-whisper inspects the first thirty seconds of audio and decides the language on its own. This is convenient for multilingual teams or when loading an unfamiliar audio file, but transcription starts slightly slower because detection runs before decoding.

User Interface

CamoVoice features a high-contrast, dark-themed interface designed for simplicity and accessibility.

Main Window Layout

┌──────────────────────────────────────────────────────────────┐ │ [Load Audio File] 🎙 Record Button Status │ │ [Settings] "Click or hold spacebar" 🎤 Device │ ├──────────────────────────────────────────────────────────────┤ │ │ │ Transcription Output │ │ (editable text) │ │ │ ├──────────────────────────────────────────────────────────────┤ │ [Copy Text] [▶ Play] [Save As] [Undo Clear] [Clear] │ ├──────────────────────────────────────────────────────────────┤ │ [Input Audio: ████████░░] ← (appears during recording) │ ├──────────────────────────────────────────────────────────────┤ │ Text Size: ═══════●═══════ Mode: [Fast|Thinking] │ └──────────────────────────────────────────────────────────────┘

Recording Audio

There are two ways to record:

Click the Microphone Button: Click once to start, click again to stop
Hold Spacebar: Hold spacebar to record, release to stop (push-to-talk style)
A brief 200 ms delay prevents accidental recordings from quick taps
If you're typing in the text area, spacebar works normally

During recording, the microphone button pulses and fills with the accent color. When you stop, transcription begins automatically.

Global Dictation Hotkey

CamoVoice can also record and paste transcriptions into any other application — no need to have the CamoVoice window focused. Enable "Enable global hotkey (Ctrl+/ ⌃)" in Settings, then:

Place your cursor in any text field (email, chat, document, etc.).
Press Ctrl+** on Windows or **⌃ on macOS to start recording.
A small dark REC pill appears in the top-left of your primary screen with a blinking dot and a live audio-level bar so you always know the mic is hot.
Press Ctrl+** / **⌃ again to stop. CamoVoice transcribes in your chosen input language, copies the result to your clipboard, and automatically pastes it where your cursor is.

The global hotkey is disabled by default and only activates while the app is running. It is intentionally set to Ctrl+` (backtick) rather than Ctrl+Space or Ctrl+Shift+Space so it doesn't collide with macOS Spotlight, the input-source switcher, or common editor shortcuts.

Recording Duration Timer

For longer recordings, a duration timer appears after 1 minute of recording, showing your elapsed time and the mode's limit (e.g., 3:45 / 15:00). The timer changes color as you approach the limit:

Gray: Normal — plenty of time remaining
Orange: Warning — less than 1 minute remaining
Red: Critical — less than 30 seconds remaining

This helps you pace your recordings and avoid hitting the automatic cutoff unexpectedly.

Transcription Animation

When CamoVoice is processing your audio, you'll see an animated "Transcribing..." status with cycling dots (translated into your UI language), and the text area border will pulse with the accent color. This provides clear visual feedback that your audio is being processed.

Audio Level Indicator

While recording, CamoVoice displays a thin Input Audio meter below the action buttons that shows real-time audio levels:

Orange gradient bar: Shows current audio input level
Yellow peak indicator: Marks the highest audio peaks
Automatic display: Appears when recording starts
Smart hiding: Disappears 1.5 seconds after recording stops

This visual feedback confirms your microphone is working and picking up audio.

Loading Audio Files

Click Load Audio File or drag and drop a supported file onto the CamoVoice window to transcribe existing audio. CamoVoice supports:

WAV — Any sample rate or bit depth
MP3 — Decoded natively with the bundled miniaudio library (no ffmpeg required)
M4A / AAC — Decoded via bundled ffmpeg in packaged builds

The app automatically resamples to the required 16 kHz and converts stereo to mono. Files currently have a 90 MB (Fast mode) and 60 MB (Thinking mode) size limit to protect responsiveness on typical consumer laptops.

Note: M4A/AAC support in packaged builds is provided by a bundled ffmpeg binary and works fully offline after installation. MP3 support is fully offline without any external binary.

Transcription Output

The large text area displays your transcriptions. Each new recording or file appends to existing text, making it easy to build up longer documents.

Working with Text

Edit freely: Click into the text area to make corrections or additions
Copy Text: Copies all text to your clipboard (button flashes green to confirm)
▶ Play: Reads the transcription aloud using text-to-speech (see Playback Voice below)
Save As: Opens a save dialog with timestamp options (see below)
Undo Clear: Restores the most recently cleared text (grayed out when nothing to undo)
Clear: Removes all text and timestamps

Undo Clear

If you accidentally clear your text, you have two options: - Click the Undo Clear button (located to the left of Clear) - Press Ctrl+Z (Windows) or ⌘Z (macOS)

This restores both the text and any associated timestamps from the most recent clear action. The Undo Clear button is grayed out when there's nothing to restore.

If the undo itself was a mistake and you want to re-clear the restored text, press Ctrl+Shift+Z (Windows) or ⇧⌘Z (macOS). The text and timestamps remain available for undo afterward, so you can freely toggle between the two states.

Save As Options

When you click Save As (or press Ctrl+S / ⌘S), a dialog appears with options:

Include date at top of document: Adds a header with the current date
Include timestamp after each transcription: Adds a [Recorded: DD-Mon-YYYY HH:MM:SS] line after each transcription segment
Include edits to transcription: Saves the text exactly as it appears in the app, including any spelling corrections or edits you've made. When this is enabled, timestamps are disabled since edited text may no longer match the original transcription segments.

Export Formats

Format	Extension	Best For
Text File	`.txt`	Universal compatibility, email, plain text editors
Word Document	`.docx`	Professional documents, further editing in Microsoft Word
PDF Document	`.pdf`	Sharing, archiving, printing (read-only format)

Select your desired format from the "Save as type" dropdown in the file dialog.

Transcription Modes

CamoVoice International offers two transcription modes, selectable at the bottom of the window. Both modes share the same multilingual base model (~74 MB covering 99 languages). They differ only in decoding parameters — so switching modes is instant and never reloads the model.

Mode	Speed	Accuracy	Best For
Fast	★★★★★	★★★☆☆	Quick notes, short recordings, real-time feel
Thinking	★★★☆☆	★★★★★	Important transcriptions, longer recordings

Fast mode: Greedy decoding (beam_size=1) with no voice-activity detection. Great for quick voice memos, short dictation, or when you want near-instant results.
Thinking mode: Beam-search decoding (beam_size=5) with voice-activity detection. Better accuracy for complex vocabulary, accents, and noisy audio.

Recordings have a time limit (15 minutes for Fast, 10 minutes for Thinking) to protect responsiveness; recording auto-stops and transcribes at the limit.

Settings

Click the Settings button to open the settings panel.

Input Device

Select which microphone to use for recording. Options include: - Default: Uses your system's default input device - All detected input devices are listed with their audio API

Input Language

Choose the language of the audio you will transcribe, or leave it on Auto-detect. If you choose English, Spanish, German, or French, the app's interface text also switches to that language. Other languages transcribe correctly but keep the interface in English. See the Multilingual Support section above for details.

Playback Voice

Choose the voice used for the ▶ Play feature, which reads your transcription aloud.

These are system voices installed on your device, not voices bundled with CamoVoice. To add or remove voices:

Windows: Settings → Time & Language → Speech → Manage voices
macOS: System Settings → Accessibility → Spoken Content → System Voice → Manage Voices

Show Timestamps in Transcriptions

When enabled, each transcription segment displays its recording time directly in the app:

Your transcribed text here. [Recorded: 09-Jan-2026 14:32:15]

Timestamps are always stored internally — this setting controls only whether they're visible in the app. You can still include or exclude them when saving.

Keep Window Always on Top

When enabled, the CamoVoice window stays visible above other applications. Useful for dictating while reading documents or multitasking.

Enable Global Hotkey

Toggles the system-wide Ctrl+` dictation hotkey described in the Global Dictation Hotkey section.

Custom Words

CamoVoice allows you to add custom vocabulary — specialized terms, names, jargon, or unusual words. This helps the model recognise these words more accurately.

To manage custom words:

Open Settings
Click Custom Words (n) (where n is your current word count)
In the Custom Words window:
Type a word or phrase and click Add (or press Enter)
View all your current custom words in the scrollable list
Click the ✕ next to any word to remove it
Click Clear All to remove all custom words
Click Done when finished

Notes: - Custom words are suggestions that increase recognition likelihood — they are not guaranteed to be transcribed exactly - You can add up to 100 custom words - Words are case-preserved but duplicate-checked case-insensitively - Custom words are saved automatically and persist between sessions

Accessibility Features

Scalable Interface

The Text Size slider adjusts the font size throughout the entire application — transcription text, buttons, labels, status messages, and the mode selector. Slide right for larger text, left for smaller. Your preference is automatically saved.

Keyboard Shortcuts

Action	Windows	macOS
Hold to record	Spacebar (hold)	Spacebar (hold)
Global dictation toggle (system-wide)	Ctrl+`\| ⌃`
Save As	Ctrl+S	⌘S
Undo clear	Ctrl+Z	⌘Z
Redo clear (revert undo)	Ctrl+Shift+Z	⇧⌘Z
Close dialogs	Escape	Escape

Note: The Escape key closes Settings and Save As dialogs without saving changes.

How It Works (Technical)

Model: A single multilingual base model (~74 MB, 99 languages). Both Fast and Thinking modes use the same weights.
faster-whisper: CTranslate2-based implementation that's 4–6× faster than the original Whisper with lower memory usage.
CPU with int8 quantisation: No GPU required.
Audio: Recorded at 16 kHz mono (Whisper's native sample rate) via SoundDevice, processed in memory — no temporary files.
TTS: pyttsx3 drives SAPI5 on Windows and NSSpeechSynthesizer on macOS.
UI: CustomTkinter, dark theme, minimum window 640×480.

Troubleshooting

Transcription is in the wrong language

Open Settings → Input Language and pick the language you are speaking, or leave on Auto-detect for mixed-language audio.

"No voices available" for Playback

Your system may not have TTS voices installed. Install voices through your operating system's speech settings.

Recording doesn't start

Check that your microphone is connected and selected in Settings
Ensure no other application is exclusively using the microphone
Try selecting a different Input Device

Transcription is inaccurate

Speak clearly and at a moderate pace
Reduce background noise
Try switching to Thinking mode for better accuracy
Make sure Input Language matches the actual language in the audio

App is slow on first transcription

The first transcription may take slightly longer as the model initialises. Subsequent transcriptions will be faster.

Tips for Best Results

Speak clearly with natural pacing — the model handles conversational speech well
Minimise background noise — clearer audio = better results
Use Fast mode for quick notes, Thinking mode for important transcriptions
Edit as you go — freely adjust the text manually to fix mistakes or reshape text around transcriptions
Set Input Language when you know it up front — it avoids the auto-detection pass and can improve accuracy

Privacy Commitment

✓ No internet connection required after installation
✓ No accounts, no sign-ups, no authentication
✓ No analytics or usage tracking
✓ No audio sent to any server
✓ Settings stored locally in plain JSON
✓ Model bundled locally — no APIs, no third-party exposure

Guía del usuario (Español)

Reconocimiento de voz multilingüe, privado y sin conexión

Descripción general

CamoVoice International es una aplicación de reconocimiento de voz para escritorio diseñada con un principio fundamental: tu voz permanece en tu dispositivo. A diferencia de los servicios de transcripción en la nube, CamoVoice procesa todo localmente mediante un modelo de IA multilingüe incluido — sin conexión a internet, sin que los datos salgan de tu equipo y sin ninguna telemetría.

Privacidad y arquitectura

100 % sin conexión: toda la transcripción se realiza en tu equipo con faster-whisper, que funciona localmente
Cero telemetría: CamoVoice no recopila nada. Sin análisis de uso, sin informes de errores, sin muestras de audio
Sin dependencias en la nube: una vez instalado, la aplicación funciona sin conexión
Configuración local: tus preferencias se guardan en un archivo settings.json en tu dispositivo y se cargan automáticamente al abrir
Multilingüe: transcribe inglés, español, alemán, francés, italiano, portugués, neerlandés, polaco, sueco y ruso — o detecta el idioma automáticamente

Compatibilidad multilingüe

CamoVoice International admite la transcripción en diez idiomas, además de la detección automática del idioma en el audio.

Elegir un idioma

Abre Configuración y usa el desplegable Idioma de entrada. La selección tiene dos efectos:

Transcripción: el idioma elegido se pasa a faster-whisper para que sepa qué fonemas esperar. Esto suele mejorar la precisión frente a la detección automática cuando conoces el idioma de antemano.
Idioma de la interfaz: si eliges inglés, español, alemán o francés, los botones de la ventana principal, la ventana de Configuración, la ventana de Palabras personalizadas, el cuadro de diálogo de Guardar como y los mensajes de estado habituales se muestran en ese idioma. Los demás idiomas se transcriben correctamente, pero la interfaz permanece en inglés.

Detección automática

Si dejas el idioma de entrada en Detección automática, faster-whisper analiza los primeros treinta segundos de audio y determina el idioma. Es útil para equipos multilingües o para archivos desconocidos, aunque la transcripción comienza algo más lenta porque la detección se ejecuta antes de la decodificación.

Interfaz

CamoVoice usa una interfaz oscura de alto contraste pensada para la sencillez y la accesibilidad.

Grabar audio

Botón del micrófono: haz clic una vez para iniciar, haz clic de nuevo para detener.
Mantener la barra espaciadora: mantén la barra espaciadora para grabar, suéltala para detener (estilo pulsar-para-hablar).
Un retardo breve de 200 ms evita grabaciones accidentales al pulsar rápidamente.
Si estás escribiendo en el área de texto, la barra espaciadora funciona con normalidad.

Atajo global de dictado

CamoVoice también puede grabar y pegar transcripciones en cualquier otra aplicación. Activa «Activar atajo global (Ctrl+/ ⌃)» en Configuración y:

Coloca el cursor en cualquier campo de texto.
Pulsa Ctrl+** (Windows) o **⌃ (macOS) para empezar a grabar.
Aparece una pequeña píldora oscura REC en la esquina superior izquierda con un punto parpadeante y una barra de nivel de audio en directo.
Pulsa Ctrl+** / **⌃ de nuevo para detener. CamoVoice transcribe en el idioma seleccionado, copia el resultado al portapapeles y lo pega donde esté el cursor.

Cargar archivos de audio

Haz clic en Cargar archivo de audio. Se admiten WAV, MP3, M4A y AAC. Los archivos se remuestrean automáticamente a 16 kHz mono. Los límites son 90 MB (Rápido) y 60 MB (Preciso).

Trabajar con el texto

Editar libremente: haz clic en el área de texto para hacer correcciones o añadidos
Copiar texto: copia todo el texto al portapapeles
▶ Reproducir: lee la transcripción en voz alta con texto a voz
Guardar como: abre el cuadro de diálogo con opciones de marcas de tiempo
Deshacer borrado: restaura el último texto borrado (Ctrl+Z / ⌘Z). Para rehacerlo, Ctrl+Shift+Z / ⇧⌘Z.

Opciones de Guardar como

Al pulsar Guardar como (Ctrl+S / ⌘S) aparece un cuadro con opciones para incluir la fecha, incluir una marca de tiempo tras cada transcripción o incluir las ediciones manuales. Formatos de exportación: TXT, DOCX y PDF.

Modos de transcripción

Ambos modos usan el mismo modelo multilingüe base (~74 MB, 99 idiomas) y se diferencian sólo en la decodificación:

Modo	Velocidad	Precisión	Mejor para
Rápido	★★★★★	★★★☆☆	Notas rápidas, grabaciones cortas
Preciso	★★★☆☆	★★★★★	Grabaciones importantes o largas

Las grabaciones tienen un límite (15 minutos en Rápido, 10 minutos en Preciso); se detienen y transcriben automáticamente al llegar al límite.

Configuración

Dispositivo de entrada: elige el micrófono.
Idioma de entrada: elige el idioma o Detección automática. También cambia el idioma de la interfaz para en/es/de/fr.
Voz de reproducción: voces del sistema operativo (no incluidas con la app).
Mostrar marcas de tiempo en las transcripciones: muestra fecha y hora tras cada transcripción.
Mantener la ventana siempre visible: mantiene CamoVoice por encima de otras aplicaciones.
Activar atajo global (Ctrl+`): activa el dictado a nivel de sistema.
Palabras personalizadas: hasta 100 términos, nombres o jerga para mejorar el reconocimiento.

Atajos de teclado

Acción	Windows	macOS
Mantener para grabar	Espacio (mantener)	Espacio (mantener)
Atajo global de dictado	Ctrl+`\| ⌃`
Guardar como	Ctrl+S	⌘S
Deshacer borrado	Ctrl+Z	⌘Z
Rehacer borrado	Ctrl+Shift+Z	⇧⌘Z
Cerrar diálogos	Esc	Esc

Solución de problemas

La transcripción está en otro idioma: abre Configuración → Idioma de entrada y selecciona el idioma correcto o déjalo en Detección automática.
"No hay voces disponibles": instala voces a través de la configuración de voz de tu sistema operativo.
No inicia la grabación: comprueba el micrófono, que no esté bloqueado por otra aplicación, y prueba otro dispositivo de entrada.
Transcripción inexacta: habla con claridad, reduce el ruido de fondo, prueba el modo Preciso y confirma el idioma de entrada.

Compromiso de privacidad

✓ Sin conexión a internet tras la instalación
✓ Sin cuentas ni registros
✓ Sin análisis ni seguimiento
✓ Sin envío de audio a ningún servidor
✓ Configuración guardada localmente en JSON
✓ Modelo incluido localmente, sin APIs de terceros

Benutzerhandbuch (Deutsch)

Vollständig private, offline funktionierende mehrsprachige Sprache-zu-Text-Anwendung

Überblick

CamoVoice International ist eine Desktop-Sprache-zu-Text-Anwendung mit einem Kernprinzip: Ihre Stimme bleibt auf Ihrem Gerät. Im Gegensatz zu Cloud-Transkriptionsdiensten verarbeitet CamoVoice alles lokal mit einem mitgelieferten mehrsprachigen KI-Modell — keine Internetverbindung nötig, keine Daten verlassen Ihren Computer, keinerlei Telemetrie.

Datenschutz und Architektur

100 % offline: Die gesamte Transkription läuft lokal über faster-whisper
Keine Telemetrie: CamoVoice sammelt nichts. Keine Nutzungsdaten, keine Absturzberichte, keine Audiobeispiele
Keine Cloud-Abhängigkeiten: Nach der Installation arbeitet die App ohne Netzverbindung
Lokale Einstellungen: Ihre Präferenzen werden in einer settings.json-Datei auf Ihrem Gerät gespeichert
Mehrsprachig: Transkribiert Englisch, Spanisch, Deutsch, Französisch, Italienisch, Portugiesisch, Niederländisch, Polnisch, Schwedisch und Russisch — oder erkennt die Sprache automatisch

Mehrsprachigkeit

CamoVoice International unterstützt zehn Sprachen sowie automatische Spracherkennung.

Sprache wählen

Öffnen Sie Einstellungen und nutzen Sie das Dropdown-Menü Eingabesprache. Die Auswahl hat zwei Effekte:

Transkription: Die gewählte Sprache wird an faster-whisper übergeben, damit es die richtigen Phoneme erwartet. Das verbessert die Genauigkeit gegenüber der automatischen Erkennung, wenn Sie die Sprache kennen.
Oberflächensprache: Bei Englisch, Spanisch, Deutsch oder Französisch wechseln Hauptfenster-Schaltflächen, Einstellungen, Fenster für eigene Wörter, Speichern-unter-Dialog und gängige Statusmeldungen in diese Sprache. Andere Sprachen werden korrekt transkribiert, die Oberfläche bleibt jedoch auf Englisch.

Automatische Erkennung

Bei Automatisch erkennen prüft faster-whisper die ersten dreißig Sekunden des Audios und ermittelt die Sprache. Praktisch für mehrsprachige Teams oder unbekannte Dateien; die Transkription startet etwas langsamer.

Oberfläche

Aufnehmen

Mikrofon-Schaltfläche: Klicken zum Starten, erneut klicken zum Stoppen.
Leertaste halten: Leertaste halten zum Aufnehmen, loslassen zum Stoppen.
Eine kurze Verzögerung von 200 ms verhindert versehentliche Aufnahmen.
Beim Tippen im Textfeld funktioniert die Leertaste normal.

Globales Tastenkürzel

CamoVoice kann Transkriptionen auch in jede andere Anwendung einfügen. Aktivieren Sie „Globales Tastenkürzel aktivieren (Ctrl+/ ⌃)" in den Einstellungen und:

Cursor in ein beliebiges Textfeld setzen.
Ctrl+** (Windows) oder **⌃ (macOS) drücken, um die Aufnahme zu starten.
Eine kleine dunkle REC-Pille erscheint oben links mit blinkendem Punkt und Live-Pegelanzeige.
Erneut Ctrl+** / **⌃ drücken, um zu stoppen. CamoVoice transkribiert, kopiert das Ergebnis in die Zwischenablage und fügt es am Cursor ein.

Audiodateien laden

Klicken Sie auf Audiodatei laden. Unterstützt werden WAV, MP3, M4A und AAC. Dateien werden automatisch auf 16 kHz Mono konvertiert. Größenlimits: 90 MB (Schnell) und 60 MB (Präzise).

Text bearbeiten

Frei bearbeiten: In das Textfeld klicken, um Korrekturen oder Ergänzungen vorzunehmen
Text kopieren: Kopiert den gesamten Text in die Zwischenablage
▶ Abspielen: Liest die Transkription per TTS vor
Speichern unter: Öffnet den Speichern-Dialog mit Zeitstempel-Optionen
Löschen rückgängig: Stellt den zuletzt gelöschten Text wieder her (Ctrl+Z / ⌘Z). Wiederherstellen mit Ctrl+Shift+Z / ⇧⌘Z.

Speichern-unter-Optionen

Bei Speichern unter (Ctrl+S / ⌘S) erscheint ein Dialog mit Optionen: Datum am Anfang einfügen, Zeitstempel nach jeder Transkription einfügen, Bearbeitungen einbeziehen. Formate: TXT, DOCX, PDF.

Transkriptionsmodi

Beide Modi verwenden dasselbe mehrsprachige base-Modell (~74 MB, 99 Sprachen) und unterscheiden sich nur in den Dekodierungsparametern:

Modus	Geschwindigkeit	Genauigkeit	Am besten für
Schnell	★★★★★	★★★☆☆	Kurze Notizen, schnelle Aufnahmen
Präzise	★★★☆☆	★★★★★	Wichtige oder längere Aufnahmen

Aufnahmezeitlimits: 15 Minuten (Schnell), 10 Minuten (Präzise); das Aufnehmen stoppt automatisch am Limit.

Einstellungen

Eingabegerät: Mikrofon wählen.
Eingabesprache: Sprache oder Automatisch erkennen. Ändert auch die Oberflächensprache für en/es/de/fr.
Wiedergabestimme: Systemstimmen (nicht von CamoVoice bereitgestellt).
Zeitstempel in Transkriptionen anzeigen: zeigt Datum/Uhrzeit nach jeder Transkription.
Fenster immer im Vordergrund: hält CamoVoice über anderen Anwendungen.
Globales Tastenkürzel aktivieren (Ctrl+`): aktiviert systemweites Diktat.
Eigene Wörter: bis zu 100 Begriffe, Namen oder Fachausdrücke zur Erkennungsverbesserung.

Tastenkürzel

Aktion	Windows	macOS
Halten zum Aufnehmen	Leertaste (halten)	Leertaste (halten)
Globales Diktat	Ctrl+`\| ⌃`
Speichern unter	Ctrl+S	⌘S
Löschen rückgängig	Ctrl+Z	⌘Z
Wiederholen	Ctrl+Shift+Z	⇧⌘Z
Dialoge schließen	Esc	Esc

Fehlerbehebung

Falsche Sprache: Einstellungen → Eingabesprache auf die gesprochene Sprache setzen oder auf Automatisch belassen.
„Keine Stimmen verfügbar": TTS-Stimmen über die Systemeinstellungen installieren.
Aufnahme startet nicht: Mikrofon prüfen, anderes Eingabegerät testen.
Ungenaue Transkription: deutlich sprechen, Hintergrundgeräusche reduzieren, Präzise-Modus versuchen, Eingabesprache prüfen.

Datenschutzversprechen

✓ Keine Internetverbindung nach der Installation
✓ Keine Konten, keine Anmeldung
✓ Keine Analytik, kein Tracking
✓ Kein Audio an Server gesendet
✓ Einstellungen lokal in JSON
✓ Modell lokal gebündelt, keine Drittanbieter-APIs

Guide de l'utilisateur (Français)

Reconnaissance vocale multilingue, privée et hors ligne

Aperçu

CamoVoice International est une application de reconnaissance vocale de bureau conçue autour d'un principe : votre voix reste sur votre appareil. Contrairement aux services de transcription en ligne, CamoVoice traite tout localement avec un modèle d'IA multilingue intégré — aucune connexion internet, aucune donnée ne quitte votre ordinateur, aucune télémétrie.

Confidentialité et architecture

100 % hors ligne : toute la transcription se fait sur votre appareil via faster-whisper
Zéro télémétrie : CamoVoice ne collecte rien. Pas d'analyse, pas de rapport d'incident, pas d'échantillon audio
Aucune dépendance cloud : après installation, l'application fonctionne sans réseau
Paramètres locaux : vos préférences sont dans un fichier settings.json chargé automatiquement
Multilingue : transcrit anglais, espagnol, allemand, français, italien, portugais, néerlandais, polonais, suédois et russe — ou détecte la langue automatiquement

Prise en charge multilingue

CamoVoice International prend en charge dix langues, ainsi que la détection automatique de la langue à partir de l'audio.

Choisir une langue

Ouvrez Paramètres et utilisez la liste déroulante Langue d'entrée. Cette sélection :

Transcription : la langue choisie est transmise à faster-whisper pour orienter la reconnaissance phonétique. Cela améliore généralement la précision par rapport à la détection automatique.
Langue de l'interface : si vous choisissez anglais, espagnol, allemand ou français, les boutons de la fenêtre principale, la fenêtre Paramètres, la fenêtre Mots personnalisés, le dialogue Enregistrer sous et les messages d'état courants passent dans cette langue. Les autres langues transcrivent correctement mais l'interface reste en anglais.

Détection automatique

En mode Détection automatique, faster-whisper analyse les trente premières secondes de l'audio et détermine la langue. Utile pour les équipes multilingues ou les fichiers inconnus ; la transcription démarre un peu plus lentement.

Interface

Enregistrer

Bouton du microphone : un clic pour démarrer, un clic pour arrêter.
Maintenir la barre d'espace : maintenez la barre d'espace pour enregistrer, relâchez pour arrêter.
Un délai bref de 200 ms évite les enregistrements accidentels.
Lors de la saisie dans la zone de texte, la barre d'espace fonctionne normalement.

Raccourci de dictée global

CamoVoice peut également dicter dans n'importe quelle autre application. Activez « Activer le raccourci global (Ctrl+/ ⌃) » dans Paramètres, puis :

Placez le curseur dans un champ de texte.
Appuyez sur Ctrl+** (Windows) ou **⌃ (macOS) pour démarrer l'enregistrement.
Une petite pastille sombre REC apparaît en haut à gauche avec un point clignotant et un indicateur de niveau audio.
Appuyez à nouveau sur Ctrl+** / **⌃ pour arrêter. CamoVoice transcrit, copie le résultat dans le presse-papiers et le colle à l'emplacement du curseur.

Charger des fichiers audio

Cliquez sur Charger un fichier audio. Formats pris en charge : WAV, MP3, M4A et AAC. Les fichiers sont automatiquement rééchantillonnés à 16 kHz mono. Limites de taille : 90 Mo (Rapide) et 60 Mo (Précis).

Travailler avec le texte

Modifier librement : cliquez dans la zone de texte pour corriger ou ajouter
Copier le texte : copie tout le texte dans le presse-papiers
▶ Lire : lit la transcription à voix haute
Enregistrer sous : ouvre le dialogue avec options d'horodatage
Annuler l'effacement : restaure le dernier texte effacé (Ctrl+Z / ⌘Z). Rétablir : Ctrl+Shift+Z / ⇧⌘Z.

Options d'enregistrement

À Enregistrer sous (Ctrl+S / ⌘S), un dialogue propose d'inclure la date en haut du document, un horodatage après chaque transcription ou les modifications manuelles. Formats : TXT, DOCX, PDF.

Modes de transcription

Les deux modes utilisent le même modèle multilingue base (~74 Mo, 99 langues) et ne diffèrent que par les paramètres de décodage :

Mode	Vitesse	Précision	Idéal pour
Rapide	★★★★★	★★★☆☆	Notes rapides, enregistrements courts
Précis	★★★☆☆	★★★★★	Transcriptions importantes ou longues

Limites d'enregistrement : 15 minutes (Rapide), 10 minutes (Précis) ; l'enregistrement s'arrête automatiquement à la limite.

Paramètres

Périphérique d'entrée : choisissez le micro.
Langue d'entrée : la langue ou Détection automatique. Change aussi la langue de l'interface pour en/es/de/fr.
Voix de lecture : voix du système d'exploitation (non fournies par CamoVoice).
Afficher les horodatages dans les transcriptions : affiche date/heure après chaque transcription.
Garder la fenêtre toujours au premier plan : maintient CamoVoice au-dessus des autres applications.
Activer le raccourci global (Ctrl+`) : active la dictée système.
Mots personnalisés : jusqu'à 100 termes, noms ou jargon pour améliorer la reconnaissance.

Raccourcis clavier

Action	Windows	macOS
Maintenir pour enregistrer	Espace (maintenir)	Espace (maintenir)
Dictée globale	Ctrl+`\| ⌃`
Enregistrer sous	Ctrl+S	⌘S
Annuler l'effacement	Ctrl+Z	⌘Z
Rétablir	Ctrl+Shift+Z	⇧⌘Z
Fermer les dialogues	Échap	Échap

Dépannage

Transcription dans la mauvaise langue : Paramètres → Langue d'entrée, choisir la langue correcte ou laisser sur Détection automatique.
« Aucune voix disponible » : installez des voix TTS via les paramètres de voix du système.
L'enregistrement ne démarre pas : vérifiez le micro, essayez un autre périphérique d'entrée.
Transcription imprécise : articulez clairement, réduisez le bruit, essayez le mode Précis, vérifiez la langue d'entrée.

Engagement de confidentialité

✓ Aucune connexion internet requise après installation
✓ Aucun compte, aucune inscription
✓ Aucune analytique, aucun suivi
✓ Aucun audio envoyé à un serveur
✓ Paramètres stockés localement au format JSON
✓ Modèle intégré localement, sans API tierce

Guida utente (Italiano)

Riconoscimento vocale multilingue, privato e offline

Panoramica

CamoVoice International è un'applicazione desktop di riconoscimento vocale progettata intorno a un principio fondamentale: la tua voce resta sul tuo dispositivo. A differenza dei servizi di trascrizione cloud, CamoVoice elabora tutto localmente tramite un modello di IA multilingue integrato — nessuna connessione internet, nessun dato lascia il computer, nessuna telemetria.

Privacy e architettura

100 % offline: tutta la trascrizione avviene sul tuo dispositivo tramite faster-whisper
Zero telemetria: CamoVoice non raccoglie nulla. Nessuna analisi, nessun rapporto, nessun campione audio
Nessuna dipendenza cloud: dopo l'installazione, l'app funziona senza rete
Impostazioni locali: le preferenze sono in settings.json sul dispositivo
Multilingue: trascrive inglese, spagnolo, tedesco, francese, italiano, portoghese, olandese, polacco, svedese e russo — oppure rileva la lingua automaticamente

Supporto multilingue

CamoVoice International supporta dieci lingue e il rilevamento automatico della lingua.

Scegliere una lingua

Apri Impostazioni e usa il menu a tendina Lingua di ingresso. La selezione:

Trascrizione: la lingua scelta viene passata a faster-whisper per orientare il riconoscimento fonetico, migliorando di solito la precisione rispetto al rilevamento automatico.
Lingua dell'interfaccia: se scegli inglese, spagnolo, tedesco o francese, i pulsanti della finestra principale, la finestra Impostazioni, Parole personalizzate, Salva con nome e i messaggi di stato comuni passano in quella lingua. Le altre lingue trascrivono correttamente ma l'interfaccia resta in inglese.

Scegliendo Rilevamento automatico, faster-whisper analizza i primi trenta secondi dell'audio per determinare la lingua.

Interfaccia

Registrare

Pulsante del microfono: clic per iniziare, clic per fermare.
Tieni premuta la barra spaziatrice: tieni premuto per registrare, rilascia per fermare.
Un breve ritardo di 200 ms evita registrazioni accidentali.

Tasto di dettatura globale

Attiva « Abilita tasto globale (Ctrl+/ ⌃) » in Impostazioni. Premi Ctrl+/ ⌃ per avviare e fermare; CamoVoice incolla la trascrizione nel punto del cursore.

Caricare file audio

Formati: WAV, MP3, M4A, AAC. Ricampionati automaticamente a 16 kHz mono. Limiti: 90 MB (Veloce) e 60 MB (Preciso).

Lavorare con il testo

Modifica libera: clicca nell'area di testo per correggere o aggiungere
Copia testo: copia tutto negli appunti
▶ Riproduci: legge la trascrizione ad alta voce
Salva con nome: dialogo con opzioni di timestamp (data, timestamp per segmento, includere modifiche)
Annulla cancellazione: ripristina l'ultimo testo cancellato (Ctrl+Z / ⌘Z). Ripeti: Ctrl+Shift+Z / ⇧⌘Z.

Formati di esportazione: TXT, DOCX, PDF.

Modalità di trascrizione

Entrambe le modalità usano lo stesso modello multilingue base (~74 MB, 99 lingue):

Modalità	Velocità	Precisione
Veloce	★★★★★	★★★☆☆
Preciso	★★★☆☆	★★★★★

Limiti di tempo: 15 minuti (Veloce), 10 minuti (Preciso).

Impostazioni

Dispositivo di ingresso: scegli il microfono.
Lingua di ingresso: lingua o Rilevamento automatico.
Voce di riproduzione: voci del sistema operativo.
Mostra timestamp nelle trascrizioni.
Mantieni la finestra sempre in primo piano.
Abilita tasto globale (Ctrl+`).
Parole personalizzate: fino a 100 termini per migliorare il riconoscimento.

Scorciatoie da tastiera

Azione	Windows	macOS
Tenere premuto per registrare	Spazio	Spazio
Dettatura globale	Ctrl+`\| ⌃`
Salva con nome	Ctrl+S	⌘S
Annulla cancellazione	Ctrl+Z	⌘Z
Ripeti	Ctrl+Shift+Z	⇧⌘Z
Chiudi dialoghi	Esc	Esc

Risoluzione dei problemi

Trascrizione nella lingua sbagliata: Impostazioni → Lingua di ingresso.
"Nessuna voce disponibile": installa voci TTS dal sistema.
Registrazione non parte: controlla il microfono e prova un altro dispositivo.
Trascrizione imprecisa: parla chiaramente, riduci il rumore, prova Preciso.

Impegno per la privacy

✓ Nessuna connessione internet dopo l'installazione
✓ Nessun account
✓ Nessuna analisi o tracciamento
✓ Nessun audio inviato a server
✓ Impostazioni in JSON locale
✓ Modello integrato localmente

Guia do utilizador (Português)

Reconhecimento de voz multilingue, privado e offline

Visão geral

CamoVoice International é uma aplicação de desktop de reconhecimento de voz concebida em torno de um princípio: a sua voz fica no seu dispositivo. Ao contrário dos serviços de transcrição na nuvem, o CamoVoice processa tudo localmente através de um modelo de IA multilingue incluído — sem ligação à internet, sem que os dados saiam do computador, sem telemetria.

Privacidade e arquitetura

100 % offline: toda a transcrição é feita localmente via faster-whisper
Zero telemetria: o CamoVoice não recolhe nada. Sem análises, sem relatórios de falhas, sem amostras de áudio
Sem dependências na nuvem: após a instalação, a aplicação funciona sem rede
Definições locais: as preferências são guardadas num ficheiro settings.json
Multilingue: transcreve inglês, espanhol, alemão, francês, italiano, português, neerlandês, polaco, sueco e russo — ou deteta a língua automaticamente

Suporte multilingue

Abra Configurações e escolha no menu Idioma de entrada. A seleção determina:

Transcrição: a língua escolhida é passada ao faster-whisper para orientar o reconhecimento fonético, normalmente melhorando a precisão face à deteção automática.
Idioma da interface: se escolher inglês, espanhol, alemão ou francês, os botões da janela principal, a janela Configurações, Palavras personalizadas, Guardar como e as mensagens de estado comuns passam para essa língua. As outras línguas transcrevem corretamente mas mantêm a interface em inglês.

Com Deteção automática, o faster-whisper analisa os primeiros trinta segundos do áudio.

Interface

Gravar

Botão do microfone: clique para iniciar, clique para parar.
Manter a barra de espaço: manter para gravar, largar para parar.

Tecla de atalho global

Em Configurações ative « Ativar atalho global (Ctrl+/ ⌃) ». Prima Ctrl+/ ⌃ para começar e parar; o CamoVoice cola a transcrição na posição do cursor.

Carregar ficheiros de áudio

Formatos: WAV, MP3, M4A e AAC. Remostrados automaticamente a 16 kHz mono. Limites: 90 MB (Rápido) e 60 MB (Preciso).

Trabalhar com o texto

Editar livremente: clique na área de texto para corrigir ou acrescentar
Copiar texto: copia tudo para a área de transferência
▶ Reproduzir: lê a transcrição em voz alta
Guardar como: abre o diálogo com opções de data e marcas temporais
Anular limpeza: restaura o último texto apagado (Ctrl+Z / ⌘Z). Refazer: Ctrl+Shift+Z / ⇧⌘Z.

Formatos de exportação: TXT, DOCX, PDF.

Modos de transcrição

Ambos os modos usam o mesmo modelo multilingue base (~74 MB, 99 línguas):

Modo	Velocidade	Precisão
Rápido	★★★★★	★★★☆☆
Preciso	★★★☆☆	★★★★★

Limites: 15 min (Rápido), 10 min (Preciso).

Configurações

Dispositivo de entrada: escolher o microfone.
Idioma de entrada: língua ou Deteção automática.
Voz de reprodução: vozes do sistema operativo.
Mostrar marcas temporais nas transcrições.
Manter a janela sempre visível.
Ativar atalho global (Ctrl+`).
Palavras personalizadas: até 100 termos.

Atalhos de teclado

Ação	Windows	macOS
Manter para gravar	Espaço	Espaço
Ditado global	Ctrl+`\| ⌃`
Guardar como	Ctrl+S	⌘S
Anular limpeza	Ctrl+Z	⌘Z
Refazer	Ctrl+Shift+Z	⇧⌘Z
Fechar diálogos	Esc	Esc

Resolução de problemas

Transcrição na língua errada: Configurações → Idioma de entrada.
"Nenhuma voz disponível": instale vozes de TTS nas definições do sistema.
A gravação não começa: verifique o microfone, tente outro dispositivo.
Transcrição imprecisa: fale claramente, reduza o ruído, experimente o modo Preciso.

Compromisso de privacidade

✓ Sem ligação à internet após a instalação
✓ Sem contas
✓ Sem análise ou rastreio
✓ Sem áudio enviado a servidores
✓ Definições em JSON local
✓ Modelo incluído localmente

Gebruikershandleiding (Nederlands)

Volledig privé, offline meertalige spraak-naar-tekst

Overzicht

CamoVoice International is een desktopapplicatie voor spraak-naar-tekst met één kernprincipe: uw stem blijft op uw apparaat. In tegenstelling tot clouddiensten verwerkt CamoVoice alles lokaal met een meegeleverd meertalig AI-model — geen internetverbinding vereist, geen data verlaat uw computer, geen telemetrie.

Privacy en architectuur

100 % offline: alle transcriptie gebeurt lokaal via faster-whisper
Nul telemetrie: CamoVoice verzamelt niets
Geen cloud-afhankelijkheden: na installatie werkt de app zonder netwerk
Lokale instellingen: voorkeuren staan in settings.json op uw apparaat
Meertalig: transcribeert Engels, Spaans, Duits, Frans, Italiaans, Portugees, Nederlands, Pools, Zweeds en Russisch — of detecteert de taal automatisch

Meertalige ondersteuning

Open Instellingen en gebruik het vervolgkeuzemenu Invoertaal. Deze keuze:

Transcriptie: de gekozen taal wordt doorgegeven aan faster-whisper om de fonetische herkenning te sturen, wat doorgaans de nauwkeurigheid verbetert ten opzichte van automatische detectie.
Interfacetaal: bij Engels, Spaans, Duits of Frans schakelen de knoppen van het hoofdvenster, het venster Instellingen, Eigen woorden, Opslaan als en veelvoorkomende statusmeldingen naar die taal. Andere talen transcriberen correct, maar de interface blijft in het Engels.

Met Automatisch detecteren analyseert faster-whisper de eerste dertig seconden.

Interface

Opnemen

Microfoonknop: klik om te starten, klik om te stoppen.
Spatiebalk ingedrukt houden: ingedrukt houden om op te nemen, loslaten om te stoppen.

Globale dicteersneltoets

Activeer « Globale sneltoets inschakelen (Ctrl+/ ⌃) » in Instellingen. Druk op Ctrl+/ ⌃ om te starten en te stoppen; CamoVoice plakt de transcriptie op de cursorpositie.

Audiobestanden laden

Formaten: WAV, MP3, M4A en AAC. Automatische herbemonstering naar 16 kHz mono. Limieten: 90 MB (Snel) en 60 MB (Nauwkeurig).

Werken met tekst

Vrij bewerken: klik in het tekstgebied
Tekst kopiëren: kopieert alles naar het klembord
▶ Afspelen: leest de transcriptie voor
Opslaan als: dialoog met tijdstempelopties
Wissen ongedaan maken: herstelt laatst gewiste tekst (Ctrl+Z / ⌘Z). Opnieuw: Ctrl+Shift+Z / ⇧⌘Z.

Exportformaten: TXT, DOCX, PDF.

Transcriptiemodi

Beide modi gebruiken hetzelfde meertalige base-model (~74 MB, 99 talen):

Modus	Snelheid	Nauwkeurigheid
Snel	★★★★★	★★★☆☆
Nauwkeurig	★★★☆☆	★★★★★

Tijdslimieten: 15 minuten (Snel), 10 minuten (Nauwkeurig).

Instellingen

Invoerapparaat: kies de microfoon.
Invoertaal: taal of Automatisch detecteren.
Afspeelstem: systeemstemmen.
Tijdstempels in transcripties tonen.
Venster altijd op de voorgrond houden.
Globale sneltoets inschakelen (Ctrl+`).
Eigen woorden: tot 100 termen.

Sneltoetsen

Actie	Windows	macOS
Vasthouden om op te nemen	Spatie	Spatie
Globaal dicteren	Ctrl+`\| ⌃`
Opslaan als	Ctrl+S	⌘S
Wissen ongedaan maken	Ctrl+Z	⌘Z
Opnieuw	Ctrl+Shift+Z	⇧⌘Z
Dialogen sluiten	Esc	Esc

Problemen oplossen

Transcriptie in verkeerde taal: Instellingen → Invoertaal.
"Geen stemmen beschikbaar": installeer TTS-stemmen via de systeemspraakinstellingen.
Opname start niet: controleer microfoon, probeer een ander invoerapparaat.
Onnauwkeurige transcriptie: spreek duidelijk, verminder ruis, probeer Nauwkeurig.

Privacybelofte

✓ Geen internet na installatie
✓ Geen accounts
✓ Geen analyses of tracking
✓ Geen audio naar servers
✓ Instellingen lokaal in JSON
✓ Model lokaal meegeleverd

Podręcznik użytkownika (Polski)

W pełni prywatne, wielojęzyczne rozpoznawanie mowy offline

Przegląd

CamoVoice International to aplikacja do rozpoznawania mowy na komputer, zbudowana wokół jednej zasady: Twój głos pozostaje na Twoim urządzeniu. W odróżnieniu od usług chmurowych CamoVoice przetwarza wszystko lokalnie, korzystając z dołączonego wielojęzycznego modelu AI — bez połączenia z internetem, bez wysyłania danych, bez telemetrii.

Prywatność i architektura

100 % offline: cała transkrypcja odbywa się lokalnie przez faster-whisper
Zero telemetrii: CamoVoice nic nie zbiera
Brak zależności od chmury: po instalacji aplikacja działa bez sieci
Ustawienia lokalne: preferencje zapisane w pliku settings.json
Wielojęzyczność: transkrybuje angielski, hiszpański, niemiecki, francuski, włoski, portugalski, niderlandzki, polski, szwedzki i rosyjski — lub wykrywa język automatycznie

Obsługa wielojęzyczna

Otwórz Ustawienia i użyj listy rozwijanej Język wejściowy. Wybór:

Transkrypcja: wybrany język trafia do faster-whisper, co zwykle zwiększa dokładność w porównaniu z wykrywaniem automatycznym.
Język interfejsu: wybór angielskiego, hiszpańskiego, niemieckiego lub francuskiego tłumaczy przyciski głównego okna, Ustawienia, Własne słowa, Zapisz jako i typowe komunikaty statusu. Inne języki są poprawnie transkrybowane, ale interfejs pozostaje po angielsku.

W trybie Wykrywanie automatyczne faster-whisper analizuje pierwsze trzydzieści sekund audio.

Interfejs

Nagrywanie

Przycisk mikrofonu: kliknięcie uruchamia, kliknięcie zatrzymuje.
Przytrzymanie spacji: przytrzymaj, aby nagrywać; puść, aby zatrzymać.

Globalny skrót dyktowania

Włącz « Włącz globalny skrót (Ctrl+/ ⌃) » w Ustawieniach. Naciśnij Ctrl+/ ⌃, aby rozpocząć i zatrzymać; CamoVoice wkleja transkrypcję w miejscu kursora.

Wczytywanie plików audio

Formaty: WAV, MP3, M4A, AAC. Automatyczne próbkowanie do 16 kHz mono. Limity: 90 MB (Szybki) i 60 MB (Precyzyjny).

Praca z tekstem

Swobodna edycja: kliknij w obszarze tekstu
Kopiuj tekst: kopiuje wszystko do schowka
▶ Odtwórz: czyta transkrypcję na głos
Zapisz jako: okno z opcjami znaczników czasu
Cofnij wyczyszczenie: przywraca ostatnio wyczyszczony tekst (Ctrl+Z / ⌘Z). Ponów: Ctrl+Shift+Z / ⇧⌘Z.

Formaty eksportu: TXT, DOCX, PDF.

Tryby transkrypcji

Oba tryby używają tego samego wielojęzycznego modelu base (~74 MB, 99 języków):

Tryb	Szybkość	Dokładność
Szybki	★★★★★	★★★☆☆
Precyzyjny	★★★☆☆	★★★★★

Limity czasu: 15 min (Szybki), 10 min (Precyzyjny).

Ustawienia

Urządzenie wejściowe: wybór mikrofonu.
Język wejściowy: język lub Wykrywanie automatyczne.
Głos odtwarzania: głosy systemu operacyjnego.
Pokaż znaczniki czasu w transkrypcjach.
Trzymaj okno zawsze na wierzchu.
Włącz globalny skrót (Ctrl+`).
Własne słowa: do 100 terminów.

Skróty klawiaturowe

Akcja	Windows	macOS
Przytrzymaj, aby nagrywać	Spacja	Spacja
Globalne dyktowanie	Ctrl+`\| ⌃`
Zapisz jako	Ctrl+S	⌘S
Cofnij wyczyszczenie	Ctrl+Z	⌘Z
Ponów	Ctrl+Shift+Z	⇧⌘Z
Zamknij okna dialogowe	Esc	Esc

Rozwiązywanie problemów

Transkrypcja w złym języku: Ustawienia → Język wejściowy.
„Brak dostępnych głosów": zainstaluj głosy TTS w ustawieniach systemu.
Nagrywanie nie startuje: sprawdź mikrofon, wypróbuj inne urządzenie.
Niedokładna transkrypcja: mów wyraźnie, zmniejsz hałas, spróbuj trybu Precyzyjnego.

Zobowiązanie prywatności

✓ Brak połączenia z internetem po instalacji
✓ Brak kont
✓ Brak analityki ani śledzenia
✓ Brak dźwięku wysyłanego na serwery
✓ Ustawienia lokalnie w JSON
✓ Model dołączony lokalnie

Användarhandbok (Svenska)

Helt privat, flerspråkig tal-till-text offline

Översikt

CamoVoice International är en skrivbordsapp för tal-till-text byggd kring en grundprincip: din röst stannar på din enhet. Till skillnad från molntjänster bearbetar CamoVoice allt lokalt med en medföljande flerspråkig AI-modell — ingen internetanslutning krävs, inga data lämnar din dator, ingen telemetri.

Integritet och arkitektur

100 % offline: all transkribering sker lokalt via faster-whisper
Ingen telemetri: CamoVoice samlar inte in något
Inga molnberoenden: efter installation fungerar appen utan nätverk
Lokala inställningar: dina preferenser sparas i settings.json
Flerspråkig: transkriberar engelska, spanska, tyska, franska, italienska, portugisiska, nederländska, polska, svenska och ryska — eller detekterar språket automatiskt

Flerspråkigt stöd

Öppna Inställningar och använd rullgardinsmenyn Inmatningsspråk. Valet styr:

Transkribering: det valda språket skickas till faster-whisper för att styra den fonetiska igenkänningen, vilket oftast ger högre träffsäkerhet än automatisk detektering.
Gränssnittsspråk: vid val av engelska, spanska, tyska eller franska översätts huvudfönstrets knappar, Inställningar, Egna ord, Spara som och vanliga statusmeddelanden till det språket. Övriga språk transkriberar korrekt men gränssnittet är på engelska.

Med Automatisk identifiering analyserar faster-whisper de första trettio sekunderna.

Gränssnitt

Spela in

Mikrofonknappen: ett klick för att starta, ett klick för att stoppa.
Håll ner mellanslag: håll för att spela in, släpp för att stoppa.

Global diktatgenväg

Aktivera « Aktivera global genväg (Ctrl+/ ⌃) » i Inställningar. Tryck på Ctrl+/ ⌃ för att starta och stoppa; CamoVoice klistrar in transkriberingen vid markörpositionen.

Läsa in ljudfiler

Format: WAV, MP3, M4A och AAC. Automatisk omsampling till 16 kHz mono. Gränser: 90 MB (Snabb) och 60 MB (Noggrann).

Arbeta med texten

Redigera fritt: klicka i textfältet
Kopiera text: kopierar allt till urklipp
▶ Spela upp: läser upp transkriberingen
Spara som: dialog med tidsstämpelalternativ
Ångra rensning: återställer senast rensad text (Ctrl+Z / ⌘Z). Gör om: Ctrl+Shift+Z / ⇧⌘Z.

Exportformat: TXT, DOCX, PDF.

Transkriberingslägen

Båda lägena använder samma flerspråkiga base-modell (~74 MB, 99 språk):

Läge	Hastighet	Noggrannhet
Snabb	★★★★★	★★★☆☆
Noggrann	★★★☆☆	★★★★★

Tidsgränser: 15 min (Snabb), 10 min (Noggrann).

Inställningar

Inmatningsenhet: välj mikrofon.
Inmatningsspråk: språk eller Automatisk identifiering.
Uppspelningsröst: systemröster.
Visa tidsstämplar i transkriberingar.
Håll fönstret alltid överst.
Aktivera global genväg (Ctrl+`).
Egna ord: upp till 100 termer.

Kortkommandon

Åtgärd	Windows	macOS
Håll för inspelning	Mellanslag	Mellanslag
Global diktering	Ctrl+`\| ⌃`
Spara som	Ctrl+S	⌘S
Ångra rensning	Ctrl+Z	⌘Z
Gör om	Ctrl+Shift+Z	⇧⌘Z
Stäng dialoger	Esc	Esc

Felsökning

Transkribering på fel språk: Inställningar → Inmatningsspråk.
"Inga röster tillgängliga": installera TTS-röster via systemets röstinställningar.
Inspelning startar inte: kontrollera mikrofonen, prova en annan inmatningsenhet.
Oprecis transkribering: tala tydligt, minska bakgrundsbrus, prova Noggrann.

Integritetsåtagande

✓ Ingen internetanslutning krävs efter installation
✓ Inga konton
✓ Ingen analys eller spårning
✓ Inget ljud skickas till servrar
✓ Inställningar lokalt i JSON
✓ Modell medföljer lokalt

Руководство пользователя (Русский)

Полностью приватное, многоязычное распознавание речи без подключения к сети

Обзор

CamoVoice International — это настольное приложение распознавания речи, построенное вокруг одного принципа: ваш голос остаётся на вашем устройстве. В отличие от облачных сервисов, CamoVoice обрабатывает всё локально с помощью встроенной многоязычной ИИ-модели — без интернета, без передачи данных, без телеметрии.

Конфиденциальность и архитектура

100 % офлайн: вся расшифровка выполняется локально через faster-whisper
Без телеметрии: CamoVoice ничего не собирает
Без облачных зависимостей: после установки приложение работает без сети
Локальные настройки: параметры хранятся в файле settings.json
Многоязычность: расшифровывает английский, испанский, немецкий, французский, итальянский, португальский, нидерландский, польский, шведский и русский — или определяет язык автоматически

Поддержка языков

Откройте Настройки и выберите Язык ввода. Выбор определяет:

Расшифровка: выбранный язык передаётся faster-whisper для настройки фонетической модели, что обычно повышает точность по сравнению с автоматическим определением.
Язык интерфейса: при выборе английского, испанского, немецкого или французского кнопки главного окна, окно Настройки, Пользовательские слова, диалог Сохранить как и типичные сообщения статуса переводятся на этот язык. Остальные языки расшифровываются корректно, но интерфейс остаётся на английском.

В режиме Автоопределение faster-whisper анализирует первые тридцать секунд аудио.

Интерфейс

Запись

Кнопка микрофона: клик — начать, клик — остановить.
Удержание пробела: удерживайте, чтобы записывать; отпустите, чтобы остановить.

Глобальная горячая клавиша диктовки

Включите « Включить глобальную горячую клавишу (Ctrl+/ ⌃) » в Настройках. Нажмите Ctrl+/ ⌃, чтобы начать и остановить; CamoVoice вставит расшифровку в месте курсора.

Загрузка аудиофайлов

Форматы: WAV, MP3, M4A и AAC. Автоматическая передискретизация до 16 кГц моно. Лимиты: 90 МБ (Быстрый) и 60 МБ (Точный).

Работа с текстом

Свободное редактирование: щёлкните по текстовой области
Копировать текст: копирует всё в буфер обмена
▶ Воспроизвести: читает расшифровку вслух
Сохранить как: диалог с настройками меток времени
Отменить очистку: восстанавливает последний очищенный текст (Ctrl+Z / ⌘Z). Повтор: Ctrl+Shift+Z / ⇧⌘Z.

Форматы экспорта: TXT, DOCX, PDF.

Режимы расшифровки

Оба режима используют одну и ту же многоязычную модель base (~74 МБ, 99 языков):

Режим	Скорость	Точность
Быстрый	★★★★★	★★★☆☆
Точный	★★★☆☆	★★★★★

Ограничения по времени: 15 мин (Быстрый), 10 мин (Точный).

Настройки

Устройство ввода: выбор микрофона.
Язык ввода: язык или Автоопределение.
Голос воспроизведения: голоса операционной системы.
Показывать метки времени в расшифровках.
Держать окно поверх других.
Включить глобальную горячую клавишу (Ctrl+`).
Пользовательские слова: до 100 терминов.

Горячие клавиши

Действие	Windows	macOS
Удерживать для записи	Пробел	Пробел
Глобальная диктовка	Ctrl+`\| ⌃`
Сохранить как	Ctrl+S	⌘S
Отменить очистку	Ctrl+Z	⌘Z
Повторить	Ctrl+Shift+Z	⇧⌘Z
Закрыть диалоги	Esc	Esc

Устранение неполадок

Расшифровка на неправильном языке: Настройки → Язык ввода.
«Голоса недоступны»: установите TTS-голоса через настройки речи ОС.
Запись не начинается: проверьте микрофон, попробуйте другое устройство ввода.
Неточная расшифровка: говорите чётко, уменьшите фоновый шум, попробуйте режим Точный.

Обязательство о конфиденциальности

✓ Интернет не требуется после установки
✓ Без учётных записей
✓ Без аналитики и отслеживания
✓ Без отправки аудио на серверы
✓ Настройки локально в JSON
✓ Модель встроена локально

CamoVoice International v1.0.0