AI Video Prompts for Selfie Vlog Videos

Selfie vlogs are the native language of TikTok and Reels: someone holds a phone at arm’s length and talks to you like a friend while their day happens behind them. Now that video models generate synced speech, you can make one without filming yourself. Creators use the format for travel diaries, day-in-my-life clips, storytime, food reactions and small-business behind-the-scenes, and it suits faceless channels that need a human-feeling host.

The camera part is harder than it looks. Most models default to a third-person shot, as if a camera operator were standing in front of the person. A real selfie is different: the lens is inches from the face, slightly above it, the arm is in the frame, the background swings around as the vlogger walks, and their eyes keep flicking to the screen instead of the lens. The voice is close and a little breathy, with wind or room noise underneath.

This page focuses on realistic human vloggers. If you want the Bigfoot or yeti-style version, see the character vlog prompts, which use the same camera with an impossible host.

Why Selfie Vlogs Need Their Own Prompt Approach

Five details make a clip read as “someone filmed themselves”:

• Arm’s-length geometry. Front camera held slightly above eye level, face close and a little off-center, shoulder and forearm visible at the frame edge. A selfie stick widens the view and raises the angle, showing more background.
• Front-camera image quality. Wide lens that slightly enlarges whatever is nearest (nose, hand), exposure set for the face so bright skies blow out, noisy shadows in low light, and the phone screen glowing on the face at night.
• Eye line. Vloggers look at themselves on the screen, just below the lens, then glance up into the lens for emphasis. Constant perfect eye contact looks like a newsreader.
• Walk-and-talk motion. A gentle bob with each step, small horizon tilts, the background rotating when they turn to show something.
• Close, unpolished audio. The voice dominates because the mic is close, with wind hits, breathing on hikes, and ambient sound of wherever they are.

The Selfie Vlog Prompt Formula

Vertical selfie vlog video, [person description] holds the phone at arm's length with the front camera, slightly above eye level, arm visible at the frame edge.
[Location + natural light], background [moving / busy / messy real details].
[Motion: walking bob / standing still / turning to show something].
They talk to the camera [tone]: "[one or two short sentences]"
Eyes glance between the screen and the lens. Visible skin texture, no beauty filter.
Audio: voice close to the mic, [ambient], [wind / breathing / room sound]. [8-10 seconds, no subtitles].

Ready-to-Use Selfie Vlog Prompt Templates

1. Day-in-My-Life Morning Opener

The first clip of almost every lifestyle vlog.

Vertical selfie vlog video, a woman in her late twenties with messy
bun and oversized grey t-shirt holds the phone at arm's length with the
front camera in a small sunlit apartment kitchen, slightly above eye
level, forearm visible at the frame edge. She sips coffee, then talks
to the camera sleepily: "Okay. Day one of actually waking up early.
It's six fifty. I hate it." Eyes glance between the screen and the
lens. Morning window light, blown-out window behind her, visible skin
texture. Audio: kettle clicking off, her voice close to the mic.
8 seconds, no subtitles.

Why it works: A time, a small confession and a blown-out window are the realistic texture that polished AI clips usually lack.

2. Just Landed: Airport Walk-and-Talk

Travel vloggers live in this shot.

Vertical selfie vlog video, a young man with a backpack and short curly
hair walks through a bright airport arrivals hall holding the phone at
arm's length with the front camera, walking bob with each step, the
hall rotating behind him as he turns. He talks excitedly: "We made it
to Lisbon. Twelve hours, zero sleep, let's go get a pastel de nata."
Harsh overhead light, other travelers passing behind, slight motion
blur. Audio: rolling suitcases, a boarding announcement echoing, his
voice close to the mic. 8 seconds, no subtitles.

3. Summit Reveal, Out of Breath

Breathing is audio realism most prompts forget.

Vertical selfie vlog video on a rocky mountain summit on a windy clear
day. A woman in a red windbreaker holds the phone on a short selfie
stick, face close, cheeks flushed, breathing hard. She says between
breaths: "Four hours. Worth it. Look at this." She slowly turns so the
camera swings around to show a huge valley and layers of peaks behind
her, then turns back grinning. Sky blown out behind her face, deep
focus. Audio: heavy wind hits on the mic, her breathing, a distant
bird. 10 seconds, no subtitles.

Why it works: The turn-to-reveal keeps the camera in selfie position while still showing the view, and wind hits on the mic sell the altitude.

4. Caught in the Rain

Imperfect weather creates the most relatable vlog moments.

Vertical selfie vlog video on a city street in sudden heavy rain. A
teenage girl with wet hair plastered to her forehead holds the phone at
arm's length under a shop awning, laughing. She says: "The forecast
said zero percent. Zero." Raindrops on the lens, gray diffused light,
people running past behind her with bags over their heads. Handheld
wobble, eyes on the screen. Audio: loud rain on the awning, splashing
footsteps, her laughter close to the mic. 8 seconds, no subtitles.

5. Parked-Car Storytime

The car is the most common storytime studio. Keep the car parked.

Vertical selfie vlog video inside a parked car in a supermarket parking
lot, afternoon. A man in his thirties with a beard and a blue work shirt
holds the phone slightly above eye level, seatbelt off, groceries on
the passenger seat. He leans in and says in a low, dramatic voice: "So
the lady at the checkout knew my name. I have never seen her before."
Window light on one side of his face, car interior slightly dark,
visible skin texture. Audio: muffled parking lot sounds, his voice
close and quiet. 8 seconds, no subtitles.

6. Small Business Packing Orders

Behind-the-scenes clips for makers and shop owners.

Vertical selfie vlog video in a cluttered home studio with shelves of
candles. A woman in her forties with reading glasses on her head props
the phone on a stack of boxes, front camera facing her at a slightly low
angle, and talks while taping a package: "Order number three hundred.
Three hundred. I'm going to cry." Warm lamp light plus daylight from a
window, messy real background, tape and tissue paper everywhere. Audio:
tape dispenser screech, her voice slightly further from the mic. 9
seconds, no subtitles.

Why it works: A propped phone changes the angle and the audio distance, a detail that makes the vlog feel handmade. Use your own real products as image references rather than generating a fake brand.

7. Can’t Sleep, 2 A.M. Whisper Vlog

Front-camera low light has its own unmistakable look.

Vertical selfie vlog video at night in a dark bedroom, lying in bed.
A young man's face is lit only by the cool glow of the phone screen,
heavy noise in the shadows, pillow and blanket edge visible. He
whispers to the camera: "It's two a.m. and I just remembered something
I said in 2016." He stares at the lens, then covers his face with the
blanket. Audio: quiet room hum, a fan, his close whisper. 8 seconds,
no subtitles.

8. Three-Clip Mini Vlog (Multi-Shot)

For models that generate several connected shots in one go.

Shot 1: Vertical selfie vlog, a young woman in a yellow raincoat walks
along a harbor at sunrise holding the phone at arm's length, saying:
"Morning. We're taking the ferry to the island."
Shot 2: Same woman, same raincoat, selfie angle on the ferry deck,
wind whipping her hair, shouting over the engine: "It's so windy!"
Shot 3: Same woman on a quiet island path, selfie angle, smiling and
out of breath: "Okay, this was a good idea."
Natural light throughout, visible skin texture, walking bob, voice
close to the mic with wind and engine sounds. No subtitles.

Model Tips for Selfie Vlogs

Veo 3.1 is a natural fit for talk-to-camera clips. It generates speech, lip sync and ambience together, supports vertical 9:16, and its video extension feature continues a clip from where it ended, which suits a vlog that keeps going. Clips run up to 8 seconds, so write one thought per clip. See the Veo 3 prompt generator.

Kling 3.0 can generate up to six connected shots in one session while keeping the character consistent, which fits the three-clip mini vlog above. It lip-syncs English, Chinese, Japanese, Korean and Spanish. Try the Kling AI prompt generator.

Seedance 2.0 takes image, video and audio references. A reference photo keeps a recurring vlogger’s face the same from clip to clip. For more on this, read the guide to consistent characters in AI video.

Hailuo 3.0 outputs 2K with stereo audio and up to 15 seconds, enough for a longer storytime line. The Hailuo prompt generator page covers its prompt style.

Common Selfie Vlog Prompt Mistakes

• Third-person framing. If you do not say the person is holding the phone, you get a camera operator. Always describe the arm.
• Staring into the lens nonstop. Real vloggers watch themselves on screen. Ask for glances.
• Beauty-filter skin. Poreless faces are the fastest AI tell. Ask for visible skin texture.
• Monologues. Two sentences per clip. Cut long thoughts into several clips, the way vloggers jump-cut.
• Driving while vlogging. Keep car clips parked. A prompt that shows someone filming while driving normalizes something dangerous and platforms may restrict it.

Label It as AI

A realistic selfie vlog shows a person who does not exist speaking as if they do. Label it as AI in the caption and with the platform’s disclosure setting, never mimic a real creator’s face, voice or catchphrases, and never use a generated vlogger to give “honest reviews” of products they never used.

How LzyPrompt Generates These

A good vlog needs lots of short clips, and each one needs the arm, the eye line, the light and the audio written in. With LzyPrompt, type the scenario (“solo traveler’s first day in Istanbul”), set Camera Style to selfie shot or keep auto-detect, and you get 10 variations covering different moments, locations and moods, each with a scene title, the full prompt and a caption. That is close to a whole vlog storyboard in one run. Export all 10 to CSV and generate them back to back.

Try LzyPrompt free for 7 days and plan your next AI vlog in minutes.