Every feature, start to finish.

1. First run

On first launch Cuefyr downloads the two English speech packs it needs to follow your voice. That is a one-time download, with a progress bar, and it needs a connection. Nothing else about reading a script does.

After that the app opens on Scripts, with the left sidebar holding the three places you work: Scripts, Reports, Settings, plus this Guide.

Give the microphone permission the first time you start a read. On macOS you also grant camera permission the first time you record video, and Screen and System Audio is never requested because Cuefyr does not capture your screen.

2. Scripts

The sidebar lists every script. The buttons at the top of the list create a new script, choose which folder scripts live in, and open that folder in Explorer or Finder.

• New makes an empty script you can name and type into.
• Drop a .txt file onto the editor to import it.
• Rename by editing the title above the editor, then clicking away.
• Scripts are plain .txt files in your scripts folder. Edit them in any other editor and Cuefyr picks up the change.

Open prompter starts the reading view with the script you have selected.

3. Writing a script

Type or paste plain text. Two things get special treatment.

Square brackets are stage directions

Anything in [brackets] shows on screen dimmed, is never counted as a word to speak, and is skipped by the voice tracker. Use them for what your body does:

[smile] Good morning, everyone.
[step forward] This is the part that matters.

Enhance

Enhance inserts those directions for you. It reads the script and adds one physical cue per line where the words invite one: a question gets a lean-in, an exclamation a step forward, imagery a gesture, a long line a breath, the close eye contact. It runs on-device, uses no AI service, and is fully reversible: the button turns into Plain text, which strips only the cues it inserted and leaves any brackets you wrote yourself alone.

Line breaks matter. Each line is one segment in the report, so break the script the way you would breathe it.

4. Reading: the prompter

The prompter takes over the window. The script sits in the middle, the control panel on the right.

• Start begins a read with a 3, 2, 1 countdown, then the script starts following your voice. It becomes Stop while reading.
• Voice / Video above Start decides what the take records: the microphone alone, or camera and microphone. Locked while a take is running.
• Pause holds everything, the tracking and the recording. Resume picks up where you stopped.
• Jump to: Top returns to the first line. Paragraph returns to the start of the paragraph you are in, which is the fastest fix when you want to run a section again.
• Float puts the window over other apps. See section 10.
• Settings opens the Setup panel for look and coaching. See section 7.
• Reports leaves the prompter and opens the report for the read you just finished.

A thin progress line along the top edge shows how far through the script you are.

5. Voice sync and manual scroll

Cuefyr holds the word you are speaking on the read line and scrolls the page under it. It expects you to be human about it:

• Pause as long as you like. The script waits. It does not drift on.
• Repeat a line, or restart a sentence, and it re-anchors on where you actually are rather than where it thought you were.
• Skip ahead and it catches up when it recognises the words you jumped to.
• Brackets are never expected to be spoken, so a [pause] does not stall the tracker.

Take manual control at any time with the arrow keys, the mouse wheel, or a trackpad. A quiet Manual chip appears while you are scrolling by hand. Press Esc, or click that chip, to hand the line back to your voice.

6. Keys and clickers

A presenter clicker works: most send the same key as Space or the arrows, so start, stop, and line stepping all work from the remote.

The last two are global shortcuts. They fire while another app has focus, which is the point: you are on a call, Zoom has the keyboard, and you still need to get out of overlay.

7. Look: font, opacity, mirror

Settings in the prompter panel opens Setup, which has two tabs.

Look

The settings that change while you read:

• Font size of the script.
• Opacity of the background behind the text, from a solid box down to nothing, which is what you want when floating over another app.
• Line position of the read line, from near the top of the screen to near the bottom. Put it just under your camera so your eyes stay in the lens.
• Mirror flips the text for a beam-splitter teleprompter rig.
• Colours for the spoken words, the bracket cues, and the background.

Every one of these persists across restarts.

Coaching

• Delivery cues Pro render the verbal layer as typography on the script itself: the power word grows, a slow line spreads out, a pause opens a gap, a pitch change shows as an arrow. Physical direction stays in your [brackets].
• Target length as mm:ss. Set 0:00 to turn it off.

8. Timer and target pace

The clock sits at the top right of the reading area and runs while you read.

Set a target in Setup and the clock also shows whether you are on pace, comparing how much of the script you have read against how much of your time you have spent. Green means on pace, amber behind, blue ahead. A five percent band either side counts as on pace, so the dot does not flicker at every sentence.

9. Recording a take

Choose Voice or Video above Start, then Start. The take records for as long as the read runs.

• Voice records the microphone only. The camera light stays off and no self-view appears.
• Video records camera and microphone, and shows a small self-view you can drag anywhere on screen so you can frame yourself. The self-view is mirrored, like a mirror, and is never part of the recorded file.
• 16:9 / 9:16 sets the shape of the recorded video, landscape for a webinar, portrait for social. You can switch it mid-take.
• A red Live chip shows while recording. Click it to stop.
• Pause holds the recording as well as the tracking.

Recordings are saved to your recordings folder. The report for that read links to the file, which is what makes the film room work.

Pick which camera and microphone to use in Settings, Devices. If the lists are empty, run one recording first so the browser engine can see your devices, then come back.

10. Floating over Zoom, Teams, Meet

Float (or Ctrl + Shift + O) drops the window into overlay mode: no title bar, no control bar, just the script floating above whatever else is on screen. Drag the strip at the top left to move it, the corner handle to resize it, and use the opacity slider in the float strip to go from a backing wash down to bare text over your call.

Click-through (Ctrl + Shift + T) makes the window ignore your cursor entirely, so clicks land on the app underneath. You can drive the Zoom controls while the script floats over them.

Keep window on top of other apps in Settings, General does the same floating without hiding the chrome. It is off by default. Overlay floats regardless of that setting.

11. Reports

Every read produces a report, saved to disk, listed under Reports.

At the top, a trend bar chart of your recent reads of the same script: words per minute, with the score above each bar, so you can see whether the rehearsal is working.

Inside a report:

• Headline: a delivery score out of 100, a one-sentence verdict, and the five foundations scored out of five: pace, pause, pitch, volume, tonality. Pitch, volume, and tonality need a recorded take to measure.
• Benchmarks, where your pace and pauses sit against the ranges good delivery tends to fall in.
• Flagged lines: rushed, dragged, and skipped beats, each clickable.
• Pace across the read, a curve of speed over time, so you can see the section where you sped up.
• Segment detail, line by line.
• Language: filler words, hedging ("I think", "sort of"), crutch words, and word variety. Only words you said that are not in the script count, so the script's own "so" or "right" is never held against you.
• Delivery shape and Coach's notes, the plain-language version of everything above.
• On camera Pro when the take was recorded with video: eye contact, framing, and movement, analysed on-device after the take. A progress bar shows the analysis running.
• Versus your average, once you have enough reads to compare against.

Download PDF exports the whole report, including the charts, as a light-on-white A4 PDF you can send to someone. Reveal folder opens the folder where reports are stored.

12. Film room

If the read was recorded, the report has a film room: your take playing with the script line that belongs under the playhead shown beneath the video, tagged when that line was rushed or dragged.

Click any flagged line in the report to open the film room at that moment. That is the loop the whole app is built for: the report says you rushed the close, one click and you are watching yourself rush the close.

13. Languages

Settings, General, Script language picks English, Hindi, or Urdu. English is ready after first run. Hindi and Urdu are separate downloads, offered right there in Settings with a progress readout and a Cancel button, and a "Check for updates" button afterwards.

A language change applies the next time you open a script in the prompter, not mid-read.

14. Free and Pro

The teleprompter is free and stays free: scripts, voice-synced scroll, manual scroll, the look controls, overlay and click-through, the timer, and recording.

A licence unlocks the coaching layer: the on-camera analysis, the voice analysis, and the delivery cues on the script.

Paste your key in Settings, Licence and press Activate. The chip at the bottom of the sidebar shows Free or Pro. Change key swaps it, and a bad key never overwrites a working one. Remove key drops this machine back to Free, behind a confirm.

See pricing for what each tier costs.

15. Files and folders

Three folders, each changeable in Settings, Storage, each with an Open button that reveals it in Explorer or Finder:

They default to the app's data folder. Point them at Dropbox, Drive, or a project folder if you would rather keep everything with the rest of the work.

16. Updates

Cuefyr checks for a new version once per launch, quietly. If there is one, a banner offers it. You can also check any time in Settings, Updates, which shows your version, the check button, and the version history of what changed in each release.

Downloads show progress and can be cancelled. The update installs when you restart, and the app offers to do that for you.

17. Privacy

• Speech recognition runs on your machine with a local model. Audio is never uploaded.
• Recordings and reports are files on your disk, in folders you choose.
• The on-camera and voice analyses run on-device after the take.
• The only network calls Cuefyr makes are the speech pack downloads, the language pack downloads, and the update check.
• There is no account, no sign-in, and no telemetry.

18. Troubleshooting

The script does not follow me. Check the microphone in Settings, Devices, and that the level moves when you speak. Then check you actually pressed Start and the countdown finished. If you scrolled by hand, the Manual chip is showing and voice is not driving the line: press Esc.

It follows me but lands a word or two late. That is the tracker being deliberately smooth rather than twitchy. Put the read line higher on screen so you have more runway above it.

It jumps to the wrong place. Repeated phrases in a script can look identical to a tracker. Read on and it re-anchors, or press Paragraph to reset to the top of the section.

Names and unusual words are missed. Speech models guess at rare words. Cuefyr already rescues names that sound close to the script. Extremely unusual spellings can still miss; the tracker recovers on the next few words.

No devices in Settings, Devices. Record once so the app can ask for permission, then the lists fill.

A report says the recording could not be found. The video was moved, renamed, or deleted after the read, or the recordings folder was changed. The report itself is unaffected; only the film room needs the file.

Click-through left me unable to click Cuefyr. That is what it does. Press Ctrl + Shift + T, or use the tray icon.

Windows SmartScreen or macOS Gatekeeper warns on install. The current builds are not code-signed yet. On Windows choose More info, then Run anyway. On macOS right-click the app and choose Open.