Teleprompter for Recorded Q&A and AMA Videos: How to Answer Audience Questions on Camera Without Sounding Scripted (iPhone)

"Answering your questions" is one of the most reliable video formats there is: the audience wrote the brief. It is also the format where reading from a script is most obvious, because a Q&A is supposed to feel like a conversation. This post is a how-to for the recorded version: a Reel, Short, TikTok, YouTube community Q&A or LinkedIn video where you collect questions first, then answer them on camera in one sitting, using the free Teleprompter: Camera Overlay on iPhone. The app records video; it does not stream, so a genuine live AMA is a different animal, covered at the end. The previous post promised Q&A where the questions come from the audience; here they do, just a day or two earlier.

Step 1: collect the questions where your audience already talks

Three sources cover almost everyone:

• Comments on a recent post or video. Ask directly: "What do you want to know about X? I'll answer the best ones on camera this week." Comments are public, so other people upvote the questions they share.
• Story question stickers on Instagram or similar. Private, low effort for the asker, lots of short questions quickly. Screenshot the ones you like before the story expires.
• DMs and emails you have already received. The question someone sent privately is usually the one ten others have but did not ask.

Paste every candidate into one note, verbatim, with the asker's first name or handle if you plan to credit them. Do not tidy the wording; the original phrasing is what makes the video sound like a real exchange.

Step 2: pick 5 to 8

Five to eight questions is enough for a satisfying long version and plenty of clips. Four tests: Was it asked more than once? Can you answer it in under 60 seconds? Does the answer show something you know rather than something you sell? Would the clip make sense to someone who did not ask it? Drop anything that needs a whiteboard or a lookup. Open with an easy, broad one, put the strongest answer in the middle, and let the last one tee up the next video. Merge near-duplicates and say so: "A few of you asked this in different ways."

Step 3: script answers that still sound spontaneous

Script the structure, not the sentences. In the app, create one script for the whole session and write each question as its own heading line, verbatim, exactly as the person typed it. Then write the answer underneath in three beats:

• Short answer. One sentence. Yes, no, it depends on one thing.
• Why. Two or three sentences of reasoning in the words you would use to a friend. Contractions, half-sentences, a "honestly" if that is how you talk.
• Example. One concrete case, ideally your own. Examples are what people remember and what makes a clip shareable.

Leave a blank line after every example: that gap is where you pause, and where the next question begins when you glance back. If an answer runs past six lines, cut the why, not the example. The rules for writing a script that does not sound written apply doubly here, because the audience knows what a real answer sounds like: they asked the question.

Step 4: read the question on camera the way a host would

The script floats over the camera preview next to the front lens, so reading the question is looking at the viewer. Three habits make it natural:

• Read the question aloud, verbatim, as a question. Let your voice go up at the end. If you credit the asker, say the name first: "Priya asks, how do you..." Reading it word for word is what a TV host does with a viewer letter; it reads as faithful, not scripted.
• Pause for a beat before the short answer. A half-second of thinking face separates a conversation from a recital. Voice-follow does not mind the silence; it waits.
• Give the short answer to the lens, not to the text. You wrote one sentence so you could say it from memory. Then drop back to the script for the why.

Voice-follow: the script waits while you ad-lib

Q&A is the format where voice-driven scrolling earns its place. The app listens with Apple's on-device speech recognition and advances the script only as you say the words. So when the example reminds you of a second example and you go off on it for twenty seconds, the text stays at the line you left; when you read the next question, it finds your place again. Stumble or restart a sentence and it re-syncs rather than running away. Recognition runs on the phone, the language is detected automatically and nothing is uploaded, so a Q&A in German or Spanish needs no settings change.

Classic auto-scroll is the better choice when the answers are tightly written and you want a fixed tempo for a very short clip, or when the room is noisy enough that speech recognition struggles. Set the speed with the slider, tap to pause between questions, drag to jump if you skip one. In either mode, turn the text size up a notch; you will look away and back more often than in a monologue.

Step 5: batch-record the whole Q&A, then cut into clips

Record the whole session in one sitting, in portrait 4K, from the front camera. Two ways work. One long take with a clear pause between answers gives natural edit points and a consistent look. Or stop and restart between questions, which leaves one take per question in the takes library and makes the clip export almost mechanical. Either way the saved video has no script on it, so every take is a clean, post-ready clip.

The payoff is one shoot, many posts: the full Q&A for YouTube or LinkedIn, and each answer as a Reel, Short or TikTok captioned with the question. Because the question is read aloud at the top of each answer, every clip stands alone without an intro.

When NOT to script a Q&A

• A genuine live AMA. If the questions arrive while you are on air, there is nothing to script, and this app records rather than streams, so it cannot overlay text on an Instagram or TikTok live anyway. The webinars and live streams post covers the realistic options: a second iPhone under the lens for the scripted open and close, and improvising the rest.
• Sensitive questions. Health, money, a public mistake, someone's grief. A scripted answer to a painful question sounds evasive even when it is kind. Three words as a cue, then look at the lens and speak.
• When reading makes it sound like PR. If the honest answer is "we got that wrong", the audience can hear a comms team in a tidy three-beat structure. For the one question that needs sincerity, take the scaffold down.
• When you know the answers cold. If you could answer each at a dinner table without thinking, put only the verbatim questions in the app and use it as a question list.

A sample 90-second, two-question script

Two questions, roughly 45 seconds each, as it would sit in the app: verbatim question lines, three-beat answers, a blank line after each:

Sam asks: "Do I need a ring light to start?"

No. A window is better than any light you can buy for under a hundred.
Ring lights put a doughnut in your eyes and flatten your face; daylight from the side gives you shape for free.
My first forty videos were shot facing the kitchen window at 9 a.m. The one I did with a ring light is the one I deleted.


Lena asks: "How long should a how-to video be?"

As long as the one thing you are teaching, and not a second longer.
People leave when they get the answer, so a 40-second clip that answers one question beats a 6-minute one that answers four.
I cut a 5-minute tripod video into three clips last month. The 38-second one about phone height outperformed the full version by a mile, because it answered the question in the title and stopped.

Read Sam's question to the lens, pause, say the first sentence from memory, then let the script carry you through the why and the example. The blank lines after "deleted" are the cue to stop, breathe and pick up Lena's question. Wander into a second window story and the script waits.

What the app does and does not do

Teleprompter: Camera Overlay is iOS-only and free: every feature works in the free version, with an occasional ad and a small one-time purchase to remove ads for good. No subscription, no watermark. Scripts are unlimited and sit beside the takes library; speech recognition and recording run 100% offline on the iPhone. It records portrait 4K with the script on the preview only, so nothing you read appears in the saved file. It does not stream, and it does not edit: cutting into clips happens in whatever editor you already use.

Next in this series: using a teleprompter for course-launch and announcement videos, the "big news" video where every word has been argued over and still has to sound like you.