Getting started
- Install and open Calla Voice. The first run downloads the speech models once (a few hundred megabytes); after that it works offline. A small status bar appears with a colored badge showing the current mode.
- Check the microphone. Say
microphone check
and talk normally — the meter should sit in the green band. A headset or desk mic within a foot of your mouth works best.
- Try three commands.
open chrome
· website youtube
· scroll down
. The black console window shows exactly what Calla heard, which is the fastest way to learn what phrasing works.
- Dictate something. Click into any text box (or say
click search
), say dictation mode
, and talk. Punctuation is added for you. Say scratch that
to erase the last sentence, command mode
to stop typing.
- Install the Chrome extension (Calla Voice Browser Control) so
click sign in
, show numbers
and choose 14
work on web pages.
- Say
what can I say
any time — the list opens with the commands for whatever program is in front, and help
opens the plain-language guide.
Modes
The badge in the corner shows the mode: green = Command · blue = Dictation · orange = Spell · purple = Number · gray = Sleep. You rarely need to switch: a whole command spoken by itself runs even mid-dictation.
| Say | What happens |
|---|
dictation mode / start typing | Everything you say is typed into the current program |
command mode / stop typing | Speech is treated as commands (the default) |
spell mode / spell mode off | Letters, digits and NATO alphabet type literally — mike sierra foxtrot tango , capital b seven |
number mode / number mode off | Only numbers and NATO letters type; everything else is a command (built for Excel) |
go to sleep / wake up | Ignore everything until woken |
restart calla voice | Restarts the program (after changing settings or commands) |
quit calla voice | Exit the program |
Apps & windows
| Say | What happens |
|---|
open <app> | Opens any installed program — open chrome , open word , open calla photo |
switch to <name> | Brings that window to the front — switch to excel |
switch window | Alt+Tab to the next window |
close window · minimize window · maximize window | Window controls for the front program |
close dialog / cancel | Presses Escape — closes a popup without closing the program |
show desktop | |
full screen | F11 in most programs, the right key in Excel; say it again to exit |
website <name> | Opens a site — website amazon , website youtube (hundreds are built in) |
search <anything> | Google search in your browser — works in every mode |
talk to claude | Jumps to the Claude window and starts dictating |
taskbar 3 | Opens or focuses the third pinned taskbar icon |
press <keys> | Any key combination — press control s , press alt f4 , press enter |
export my profile / import my profile | Back up or restore your commands, words and settings (a zip in Documents) |
Chrome & browsing
| Say | What happens |
|---|
new tab · close tab · reopen tab | |
next tab / previous tab | |
go back / go forward / refresh | |
address bar | Puts the cursor in the address bar — then dictate or spell |
find <word> | Finds that word on the page |
zoom in / zoom out / reset zoom | |
click bookmark <name> | Opens one of your Chrome bookmarks by name |
Clicking things by name
On web pages this uses the Chrome extension; in every other program (Excel ribbon, Settings, dialogs, the Claude app) Calla reads Windows' accessibility tree to find the control by its visible name.
| Say | What happens |
|---|
click <text> | Clicks the button, link, tab or menu item with that text — click sign in , click 10 day . Puts the cursor in a text box by its label too: click search |
show numbers | Numbered green tags appear on everything clickable on the page |
choose 14 / click 14 / fourteen | Clicks that numbered element |
hide numbers / cancel | Removes the tags |
press it | A button ignored the click? Sends a real mouse click to the last target |
what can I click | Lists the clickable names in the front window |
| Say | What happens |
|---|
scroll down / scroll up | Add a number: scroll down five |
page down / page up | Instant — fires in a fraction of a second |
go to top / go to bottom | |
click · double click · right click | At the mouse pointer (or the grid cell) |
move mouse up 50 | Nudges the pointer up/down/left/right |
| Say | Mouse grid (Dragon style) |
|---|
mouse grid | Numbered 3×3 grid over the monitor the pointer is on |
mouse grid screen two | The grid on a specific monitor |
5 or 5 3 4 | Zooms into that cell — a run of numbers drills several levels |
mouse grid 8 5 3 4 | Opens the grid and drills straight down in one breath |
click / escape | Click the center of the cell, or close the grid |
Dictation
Speak naturally — punctuation is added automatically. Say punctuation to force it: comma
, period
, question mark
, open quote
… close quote
, new line
, new paragraph
. These commands work while dictating:
| Say | What happens |
|---|
scratch that | Erases the last thing you dictated (say it again to keep going back) |
delete last word / backspace five | |
correct that / correct that to <words> | Fixes the last phrase |
select all · copy that · cut that · paste that | |
undo that / redo that / save | |
press enter (or any key) | |
spell <letters> | Spells one word without leaving dictation — spell x y z |
type <words> | Types the words literally, even if they look like a command |
command mode | Back to commands |
Cursor & editing
| Say | What happens |
|---|
move up three / move down 5 | Moves the cursor that many lines |
move left two words / move right 4 | Words or characters |
select up 3 / select down 2 | Selects that many lines |
select line / clear line | Selects the current line — or selects and erases it |
select last word / unselect | |
select <words you dictated> | Highlights those words; the next thing you say replaces them (Select-and-Say) |
cap that / all caps that / no caps that | Retypes the selected words as Calla Voice / CALLA VOICE / calla voice, then moves the cursor to the end of the line. One breath: cap calla voice |
start of line / end of line | |
delete that | Deletes the selection |
Spell mode & number mode
| Spell mode | |
|---|
spell mode | Orange badge. Say letters, digits or NATO: m s f t , whiskey charlie romeo , capital b seven |
w c r one nine six eight at gmail dot com | Whole words like gmail and dot com type literally |
scratch that / backspace | Work here too |
spell mode off | Back to the mode you came from |
| Number mode | |
|---|
number mode | Purple badge. Switches on by itself when Excel comes to the front |
one two five · 3.5 · minus sign 80 · 5% · $1,250 | Number talk types; ordinary words are ignored so nothing stray lands in a cell |
alpha bravo / capital delta | NATO letters type as letters |
number mode off | Back to the mode you came from |
Excel
These fire only while Excel is the front window.
| Say | What happens |
|---|
column W row 10 / cell W 10 | Jumps to that cell (NATO works: column whiskey row 10 ) |
row 42 | Jumps to that row |
next cell / previous cell / next row | Tab, Shift+Tab, Enter |
first cell / last cell | Ctrl+Home / Ctrl+End |
top of the column / bottom of data | Jump to the edge of the data |
select column / select row / clear cell | |
next sheet / previous sheet | |
hide the ribbon / full screen mode | More room for the sheet |
Training words
Works like Dragon. Watch the black console window to see what was heard, then teach it.
| Say | What happens |
|---|
find word <word> → pick two → train that | Finds an existing word, picks a match, then you say it the way you normally do — the mishear is saved |
train word → spell it → save word | Adds a brand-new word (names, tickers, jargon) |
delete that | Removes the picked word from the vocabulary |
Calla Reader PDF companion app
A full PDF reader and annotator built to be run by voice. These commands work only while Calla Reader PDF is the front window — everywhere else the same words keep their normal meaning.
| Say | Reading |
|---|
next page / previous page / next page 3 | Turn pages, with an optional count |
first page / last page / page 12 | Jump |
zoom in / zoom out / actual size | |
fit width / fit page | |
rotate right / rotate left | |
continuous / single page | Scroll mode |
side panel · dark mode · full screen | View toggles |
find <words> / next match / previous match | Search |
open file · save · save as · print · print preview · close file | File |
| Say | Marking up & editing |
|---|
highlight · underline · strikethrough · squiggly | Pick a text-markup tool, then drag over the text |
sticky note · free text · ink | Notes and freehand drawing |
rectangle · ellipse · line · arrow | Shapes |
select tool / delete annotation | Back to the pointer; remove the last mark |
annotation color / stroke color | Color pickers |
sign document / manage signatures | Place a saved signature — click to place, drag to size |
insert page before / insert page after / delete page | Page management |
extract page / extract range / append pdf | Split and merge |
export images / export text / save with password / properties | |
Calla Photo companion app
A fast photo viewer and editor where every toolbar button is a phrase. Edits autosave; undo
steps back up to 25 changes.
| Say | Viewing |
|---|
next photo / previous photo / next photo 5 | Move through the folder |
first photo / last photo | |
zoom in / zoom out / actual size | |
fit window / fit width | |
toolbar · thumbnails · text descriptions · navigation arrows · dark mode · full screen | View toggles |
open image · open folder · save · save as · close image | File |
delete photo | Sends the file to the Recycle Bin and moves to the next one |
| Say | Editing |
|---|
rotate right / rotate left / flip horizontal / flip vertical | |
crop … drag … apply | Or cancel to back out |
outline / highlight / outline settings | Circle or highlight part of a picture |
eraser / eraser color / add text | Paint over an area, or drop text on the image |
resize / straighten / brightness | Edit dialogs |
grayscale · sepia · invert · enhance · sharpen · blur | Filters |
cartoon · anime · vintage · black and white · sketch · posterize | Artistic filters |
undo / redo | |
Your own commands
Say studio
to open the Command Browser and add a phrase that types text, presses keys, opens a program, opens a website, or runs anything — my email address
, sign my name
, check the weather
. Give one command several spoken names with an alias. Everything you add is backed up with export my profile
.
