Teleprompter for YouTube videos: the phone setup that works

For YouTube, a teleprompter on the phone you are filming with is enough: the script scrolls on the screen you face, the front camera records, and your eyes stay near the lens. Turn the phone sideways for long-form so it records 16:9, keep it upright for Shorts, and record long videos in sections rather than in one take.

That covers most of it. The rest is the detail that separates a video that sounds read from one that does not — the setup, the shape of the script, and the few settings that matter. The basics of running a phone prompter are in how to use a teleprompter on your phone; this page is about YouTube specifically.

Long-form or Shorts: two different setups

The two formats share a channel and almost nothing else. Decide which you are making before you set anything up, because the phone’s orientation, the script length and the way you record all follow from it.

The limits are YouTube’s own, checked October 2026: by default an account can upload videos up to 15 minutes long, and a verified one up to 12 hours or 256 GB, whichever comes first. A video counts as a Short when it is square or vertical and up to three minutes, for uploads since 15 October 2024. One condition in that page is worth knowing before you add music: a Short longer than one minute with an active Content ID claim is blocked worldwide.

The teleprompter setup for YouTube, on a phone

You do not need a glass rig for a talking-head channel. You need the phone steady, at eye height, at the right distance, with the light in front of you.

1. A tripod with a phone clamp. Hand-held is fine for a Short; a ten-minute video needs the frame not to drift.
2. Lens at eye height. Below it and the viewer looks up your nose; above it and you look up at them, and your eyes show more white.
3. About an arm’s length away, 60–90 centimetres. Further than that and the text gets hard to read; closer and the wide front lens distorts your face. Larger text lets you sit further back.
4. Light facing you. A window in front of you, slightly to one side, beats any lamp. Never a window behind you.
5. A microphone if you can. Viewers forgive soft video long before they forgive echoing audio. A clip-on mic near your chest does more for a phone video than any camera setting.

The full list, including how to frame and what to put behind you, is in talking-head video setup with a phone.

Front or back camera?

The back camera on most phones is better than the front one. But a teleprompter needs the words on the screen you are looking at, and with the back camera that screen faces away from you. Unless you have a second display or a mirror rig, record on the front camera — the gap in image quality matters far less to a viewer than eyes that drift off the lens. If you do build a rig, why teleprompters are mirrored explains the glass and the mirror setting.

How to structure a YouTube script for a teleprompter

A long-form script is read in pieces and cut together, so write it in pieces. A shape that works for most explainers:

1. Cold open (15–30 seconds). What this video answers, and why the viewer should believe you can answer it. No channel intro.
2. The promise (one or two sentences). “By the end you’ll know X, and the one mistake that makes Y fail.”
3. Sections (1–3 minutes each). One idea per section, each opening with the point and closing with a line that leads to the next.
4. The payoff. The thing you promised, said plainly.
5. The end (15–20 seconds). One ask — subscribe, the next video, a comment — and stop.

Mark the section breaks in the script with a blank line and a short heading. They become your record-and-stop points, and they are where the edit will cut. The general craft of writing for the prompter — short lines, numbers spelled the way you say them, pauses written in — is in how to write a teleprompter script.

How many words for a YouTube video?

At 150 words a minute: 750 for five minutes, 1,500 for ten, 2,250 for fifteen. Then take off 10 to 15% for pauses, b-roll and the moments you look away to demonstrate something. If you write in Hindi or Tamil the counts are lower for the same running time, for reasons covered in teleprompter speed by language; the per-minute table for every length is in how many words is a 1, 2, 3 or 5 minute speech.

Record in sections, not in one take

A ten-minute single take sounds efficient and is not. One stumble at minute eight costs you the whole thing, and you will be more tired on the retake than on the first try. Recording each section as its own take has three advantages:

• A mistake costs one section. Redo two minutes, not ten.
• The cuts are free. You were going to cut between sections anyway; a jump between two takes at a section break reads as a scene change, not an error.
• Your energy resets. Stand up, drink some water, and read the next section as if it were the first.

A practical rhythm: read a section once off-camera, record it, watch the first five seconds of the take to check focus and sound, then move on. If a take is bad, record it again straight away while the line is fresh.

And make sure a long take survives

The longer the take, the more an interruption costs. An incoming call, a low-battery shutdown or the system closing the app for memory can, in many recording apps, leave a file that will not open — not the last few seconds lost, the whole take. Why that happens, and how to test any app in four minutes, is in does a teleprompter app save your video if it crashes? Run that test before you trust a three-minute section to anything.

1080p or 4K for YouTube?

For a talking head, 1080p at 30 frames a second is enough. Most viewers watch on a phone or a laptop, YouTube re-encodes everything you upload, and a face in good light at 1080p looks sharp. 4K earns its storage in two cases:

• You crop in during the edit. A punch-in from a wide shot to a close-up for emphasis is common in talking-head editing, and at 4K it loses no visible sharpness.
• The background matters. Products on a shelf, a workspace, text you want legible.

60 frames a second matters for movement — hands, demonstrations, anything you might slow down — and much less for someone sitting and talking. HDR helps when there is a bright window in the shot that would otherwise blow out to white. All three cost storage, so check you have room before a long session.

Recording a YouTube video in Orate

Orate is a phone teleprompter that records the video itself, and it handles both YouTube setups:

1. Paste or import the script. Import takes a .txt file. For a long video, save one script per section, so each take starts at its own top. The free tier keeps three saved scripts, so a video of more than three sections needs Pro, which also adds 4K, 60fps and HDR.
2. Turn the phone for long-form. The reading band sits under the front camera and moves with it when you rotate, so in landscape the words are still next to the lens rather than across the middle of the screen.
3. Set the text and window. Text runs from 18 to 96 point; for a phone near the far end of that range, go well above the 34-point default. The reading window defaults to 45% of the screen; shorter keeps your eyes stiller.
4. Record. After the countdown (three seconds by default) the script scrolls as the recording starts. If you are running ahead or behind, drag the pace slider on the recorder; it changes speed without the script jumping.
5. Trim, then export. Trim the start and end of each take in the preview, then save or share. Export is free and unwatermarked on both tiers.

On quality: the free tier records up to 1080p at 30fps, which is what we recommend above for most talking-head videos. 4K, 60fps and HDR are the paid upgrade, and the app records at the closest quality your phone actually supports rather than failing at the countdown. Orate writes each take in two-second fragments, so if the app is killed mid-take you lose the fragment in progress, not the take — we measured that on a handset and wrote it up in what is recoverable when a recording app crashes.

For Shorts, hold the phone upright and treat it like a Reel; the scripts and word budgets in teleprompter for Reels carry straight over.

Three things that make a scripted YouTube video sound read

• Long lines. Each line is an eye movement across the screen. Bigger text and a narrower band mean shorter lines and less scanning. The geometry is in why teleprompter videos still look like you’re reading.
• An even pace. Speech speeds up and slows down; recitation does not. Write the pauses in and let the scroll be slightly slower than you think you need.
• Written grammar. “It is important to note that” is a sentence nobody says. Read the script aloud before recording and rewrite every line you would not say to a friend. More in how to sound natural when reading a script.

Questions people ask

Do YouTubers use teleprompters?

Many do for scripted talking-head videos, explainers and sponsor reads, and plenty do not. A teleprompter suits people who script word for word; people who talk from bullet points usually do better with a short outline on screen instead.

Can I use my phone as a teleprompter for YouTube?

Yes. A teleprompter app that records the video itself puts the script on the screen you face and films with the front camera. For long-form, turn the phone sideways to record 16:9; for Shorts, hold it upright.

How long can a YouTube Short be?

Up to three minutes, for square or vertical videos uploaded since 15 October 2024, according to YouTube Help (checked October 2026). A Short over one minute with an active Content ID claim is blocked.

Should I record YouTube videos in 1080p or 4K?

1080p at 30fps is enough for most talking-head channels. 4K helps if you plan to crop in for punch-ins during editing, or if you need the extra detail; it costs a lot more storage per minute.

How long a script do I need for a 10-minute YouTube video?

About 1,500 words of English at 150 words a minute; fewer for Hindi or Tamil, which Orate paces by default at 130 and 115 words a minute, a starting point rather than a measurement. Most 10-minute videos also include b-roll and pauses, so a script of 1,300 to 1,400 words is a safer target.

Should I record a long YouTube video in one take?

Rarely. Record it in sections of one to three minutes, one per part of the script, so a stumble costs one section and you can cut between them in the edit.