GoodCase Official Skill · Shared method across creators

POV and vlog presence

Create credible presence through first-person framing, handheld movement, and lived-in detail.

103 Cases · 55 Creators · AI Video · Latest example 10/21/2026

Seedance 2.5 Japanese Summer Festival Yukata Selfie Vlog

Verified installable Agent Skill

SKILL.md + .skill

Install into your agent.

npx skills add LearnPrompt/goodcase-lite --skill pov-vlog-presence --global --copy --yes --full-depth

Download .skill ↓Inspect source ↗

Method skeleton

01Define the shooter and device
02Write the movement path and incidental detail
03Keep handheld imperfection and environmental texture

How to make it

What it is for, what to prepare, and where it breaks.

Summarised by GoodCase from the representative cases below: their full prompts and our retest records. Each case keeps its creator credit and original post.

Good for

For clips that look like a real person filmed them on a phone or small camera: arm's-length selfies, talking while walking, flipping the camera to the scene and back, or POV from someone standing next to them. It is not for product ads, pure retro DV looks, or drone FPV flights.

Inputs to prepare

—A specific place and time of day
—A description of the person, or a reference image to lock face and outfit
—Three to six small events in order
—One to three lines you want spoken
—Who holds the camera, her or a friend or partner
—Aspect ratio and length

Method steps

01Decide who holds the camera. State whose hand the phone is in and how far from the face, such as an arm's-length front camera, or a boyfriend holding it a metre or so away and never appearing.
02Split the length into segments and write what happens in each by the second. The Tokyo food, gym and cycling cases all set a timeline first, then the actions.
03Write in phone-camera flaws: walking bounce, slight shake, autofocus hunting, exposure shifts between inside and outside, the occasional cropped face.
04Put the exact spoken lines in the prompt and say they are delivered to camera with matching lip movement. An opening line and a goodbye are the most common.
05Ask for diegetic sound only, list concrete ambient sounds, and say no background music and no subtitles.
06End with bans such as no cinematic moves, no drone, no slow motion, no third-person camera, no beauty filter, and repeat that face and outfit stay the same throughout.

Variables to swap

—Setting and activity: festival, side-street dessert cafe, gym, train, mountain ride
—Who films: herself at arm's length, or a friend or partner filming her
—Device: phone front camera, rear camera, action cam, handlebar mount
—Lines: an opener, a mid-clip aside, a goodbye
—Aspect and length: 9:16 vertical or 16:9, 10 to 30 seconds
—One continuous take or a multi-shot mix

Common failures and limits

—Drifting into a cinematic look. Several creators add bans for exactly this: no cinematic moves, no drone shots, no color grading, no tripod or third-person camera. (Source: prompt avoid lists)
—Face or outfit changing between shots. Most prompts lock the same face and clothes again and again; the cycling prompt also rules out a third hand. (Source: prompt constraints)
—Lines may not come out as written. In our retest of the Tokyo food case the transcript has the scripted closing line, plus an opening line the prompt never wrote. The festival retest has no transcript, so we cannot tell whether the goodbye was spoken. (Source: retest data, single generation)
—Overdoing the flaws. The late-night kitchen prompt bans excessive shake; the boyfriend POV prompt bans digital zoom and face-filling close-ups. Keep shake and focus misses light. (Source: prompt avoid lists)
—The eight-shot cycling vlog, which switches between several camera types, scored 74.4 in our single retest; the Tokyo food and festival vlogs, which stay on one phone selfie view, landed around 82. One run each, so this says nothing firm about the cause. (Source: retest data, weak evidence)

What you should get

A realistic 10 to 30 second video in whatever aspect the prompt sets; both 9:16 and 16:9 appear. Most prompts ask for diegetic sound and lip-synced lines, with no music and no subtitles. Actual length and resolution depend on the model and settings you use.

What our retests cover

These prompts were written for Seedance. On 2026-09-07 we ran four of them once each on MiniMax H3 Max 768p; all four completed and were scored by two AI judges, with no human verdict. We have not retested on Seedance itself.

Representative cases

Tokyo Street Food Phone Vlog
By @saniaspeaks_ · original post ↗
A Tokyo street-food phone vlog that times every beat from walking in to the sign-off line, then lists phone-camera flaws and bans, which makes it the most complete selfie-vlog template here.
Full prompt and retest record →
Seedance 2.5 Japanese Summer Festival Yukata Selfie Vlog
By @im_shahid7 · original post ↗
A yukata festival selfie vlog written as one short paragraph: setting, a few small events and a wave goodbye. A good minimal starting template.
Full prompt and retest record →
Seedance 30-Second Gym Day Vlog
By @ZaraIrahh · original post ↗
A gym selfie vlog that requires every shot to be her own front-camera footage, bans third-person cameras and tripods, and spells out three spoken lines with diegetic sound only.
Full prompt and retest record →
Seedance 2.5 Realistic Cycling Vlog: 30-Second Action Cam + Front Camera + Tracking Shot Edit
By @johnAGI168 · original post ↗
A cycling vlog that mixes action cam, front-camera selfie, handlebar POV and chase shots, holding continuity with rules for glasses, rising sweat and no outfit change. Useful if you want a multi-camera vlog.
Full prompt and retest record →
Seedance 2.5 Boyfriend-POV Couple Short with an Authentic Handheld Phone Look
By @liyue_ai · original post ↗
A one-take boyfriend POV where the person filming is heard but never seen, with set distances, shot-size order and reaction timing. The most detailed example of filming-someone-else POV; it has not been retested yet.
Full prompt and retest record →

Case evidence

Judge the Skill through finished work.

Showing 12 of 103

AI VideoXAdded 08/19/2026

Tokyo Street Food Phone Vlog

A highly detailed prompt for a photorealistic Tokyo food vlog that mimics phone-shot footage, complete with natural handheld shake and authentic neighborhood detail.

Skill · POV and vlog presence

Prompt

Ultra-realistic smartphone food vlog set in a charming neighborhood in Tokyo, Japan. The entire video must look like it was genuinely recorded on a modern flagship smartphone by a

Source heat · X

Stability

View case →

AI VideoX (Twitter)Added 08/19/2026

Seedance 2.5 Japanese Summer Festival Yukata Selfie Vlog

A 30-second selfie-view clip: a girl in a yukata strolls a summer festival at night, buys takoyaki, plays a scooping game, watches fireworks, and waves goodbye to the camera at the end.

Skill · POV and vlog presence

Prompt

A 30-second photorealistic smartphone travel vlog. A young Japanese woman wearing a pastel yukata explores a traditional Japanese summer festival at night. She films herself in selfie mode with natura

Source heat · X (Twitter)

Stability

View case →

AI VideoXAdded 08/19/2026

Seedance 30-Second Gym Day Vlog

Gym day: survived, sweated, and somehow left proud 💪😅 Created with Seedance 2.5 on @wavespeed_ai Prompt: Create a 30-s

Skill · POV and vlog presence

Prompt

Create a 30-second casual gym vlog featuring the same young woman in her early twenties. She has natural everyday beauty, and her face and hair remain identical throughout the entire video. Camera &…

Source heat · X

Stability

View case →

AI VideoXAdded 08/19/2026

Seedance 2.5 Realistic Cycling Vlog: 30-Second Action Cam + Front Camera + Tracking Shot Edit

A realistic 30-second vertical cycling vlog combining three camera angles: an action camera, a phone’s front camera, and tracking shots from a support vehicle. Features 4K photorealism, soft overcast light, sweaty skin with visible pores, and no beauty filters. The prompt uses structured Chinese sections such as 【Style】 and 【Duration】 to lock down camera switches and the timeline—a model example of “writing editing logic into the prompt.”

Skill · POV and vlog presence

Prompt

It really does look exactly like the real thing 😄 With a more complete 30-second narrative, Seedance 2.5 takes realism up another level 🎬 Seedance 2.5 prompt 👇 【Style】Authentic Cycling Vlog, combining an action camera + phone front ca

Source heat · X

Stability

View case →

AI VideoXAdded 08/19/2026

Seedance 2.5 Boyfriend-POV Couple Short with an Authentic Handheld Phone Look

A couples short in a phone-shot boyfriend POV style, made with Seedance 2.5. The author refined the prompt using their Seedance 2.5 director Skill so the result looks more natural, especially the blushing and expressions that earlier felt forced. The prompt specifies 20 seconds in 9:16 vertical.

Skill · POV and vlog presence

Prompt

Duration: 20 seconds Aspect ratio: 9:16 vertical Style: Couple POV with an authentic handheld-phone look (Boyfriend POV), natural summer outdoor light, one continuous take with absolutely no cuts, and the texture of a genuine casual phone r

Source heat · X

Stability

View case →

AI VideoXAdded 10/23/2026

A Child’s Day in 2000 Through Their Own Eyes

Pov : You are in 2000s Generated with Seedance 2.5 on @higgsfield Try it with prompt : STYLE RULE: First-person POV throughout, shot like a 2000 home camcorder …

Skill · POV and vlog presence

Prompt

STYLE RULE: First-person POV throughout, shot like a 2000 home camcorder (MiniDV): soft focus, slight color bleed, warm grain, 4:3 feel, orange timestamp "OCT 21 2000" in the corner. Our hands are al…

Source heat · X

Stability

View case →

AI VideoXAdded 10/19/2026

A Woman Drops Her Selfie Phone, Going from Beach to Swiss Alps

The smoothest vacation switch you’ve ever seen 🌊❄️ Created with Seedance 2.5 on @openart_ai Prompt: Ultra-realistic continuous POV selfie vlog, natural handhel…

Skill · POV and vlog presence

Prompt

Ultra-realistic continuous POV selfie vlog, natural handheld smartphone footage, one identical woman <<image throughout, preserve the exact same face and identity. Duration: 10 seconds. 0–3 seconds …

Source heat · X

Stability

View case →

AI VideoXAdded 10/19/2026

A Seoul Woman’s Morning of Shopping and Watering Plants in 2003

My kind morning routine 🙂 Made using Seedance 2.5 Prompt : Create a 30-second, ultra-realistic Korean home video from around 2003, recorded casually by a frien…

Skill · POV and vlog presence

Prompt

Create a 30-second, ultra-realistic Korean home video from around 2003, recorded casually by a friend on an early-2000s consumer DV camcorder. It must feel like raw personal footage of an ordinary m…

Source heat · X

Stability

View case →

AI VideoXAdded 10/18/2026

A Girl’s Retro Selfie Diary on Seoul’s Shopping Streets

Seedance 2.5 on @DomoAI_ Prompt CAMERA: Early-2010s smartphone front-camera selfie vlog, filmed on an old iPhone 4 or Samsung Galaxy S2. The girl holds the phon…

Skill · POV and vlog presence

Prompt

Seedance 2.5 on @DomoAI_ Prompt CAMERA: Early-2010s smartphone front-camera selfie vlog, filmed on an old iPhone 4 or Samsung Galaxy S2. The girl holds the phone at arm's length, occasionally swit…

Source heat · X

Stability

View case →

AI VideoXAdded 10/11/2026

Seedance 2.5 30-second couple bedroom POV: a boyfriend who can't comfort her, only pinch her cheek

A 30-second 16:9 photoreal slice of couple life that the author says was made with Seedance 2.5. It plays as one boyfriend-POV long take in a bedroom at night: she seizes on his line "You're quite chatty today" to tease him, he clumsily calls her cute, pinches her cheek and deadpans that her face is so soft a pinch should make the anger go away, and she finally cracks up. The prompt scripts every line, pause, micro-expression, the hand's physical path and what the mirror may never reflect. Reproduction materials are incomplete: the prompt opens by asking for an uploaded character reference image of the female lead, and the original author did not provide it; this page only includes the original prompt and the finished video. To reproduce it you need to supply your own reference image of an adult that you have the right to use and that the person has consented to. GoodCase has not retested this prompt, so reproduction results are unverified.

Skill · POV and vlog presence

Prompt

Use the uploaded character reference image as the sole reference for the female lead. She is a young adult woman. Strictly keep the same character identity throughout, and tightly lock onto the face shape, facial feature proportions, eyebro

Source heat · X

Stability

View case →

AI VideoXAdded 10/08/2026

Daily Rural Life of a Young Woman in 19th-Century Tuscany

This AI video is a Masterpiece 👏 Seedance 2.5 Prompt: 30-second ultra-realistic documentary slice-of-life video, early-2000s consumer DV camcorder aesthetic. A…

Skill · POV and vlog presence

Prompt

30-second ultra-realistic documentary slice-of-life video, early-2000s consumer DV camcorder aesthetic. A young Italian peasant woman in her early 20s, sun-weathered olive skin with freckles, dark br…

Source heat · X

Stability

View case →

AI VideoXAdded 10/05/2026

A Tokyo Girl’s Gym Day in the Early 2000s

A little old school workout session.🏋️‍♀️ Seedance 2.5 on @openart_ai Prompt Create a highly photorealistic 30-second early-2000s Japanese gym day vlog featuri…

Skill · POV and vlog presence

Prompt

A little old school workout session.🏋️‍♀️ Seedance 2.5 on @openart_ai Prompt Create a highly photorealistic 30-second early-2000s Japanese gym day vlog featuring ONE consistent young Japanese wo…

Source heat · X

Stability

View case →

Creator practice

Creator versions

Creator practices derived from this shared method, including versions not yet listed in the catalog.

saniaspeaks_ · DV Tape Korean Idol Strength04_X · Live News Broadcast ElsaSofia__AI · Multi-Ref Long Sequence AIwithkhan · Korean Alley Home Video doctorwasif · Phone Snapshot Gym Comedy johnAGI168 · Beat-Synced Fitness Ad ZaraIrahh · Selfie-Only Day-in-Life Ciri_ai’s approach · 3 cases Goodmanprotocol’s approach · 3 cases Just_sharon7’s approach · 3 cases oggii_0’s approach · 3 cases techhalla · 90s West Coast 16mm

Creator evidence

Methods stay attached to their makers.

6 Casessaniaspeaks_View creator →6 CasesStrength04_XView creator →5 CasesElsaSofia__AIView creator →4 CasesAIwithkhanView creator →4 CasesdoctorwasifView creator →4 CasesjohnAGI168View creator →4 CasesZaraIrahhView creator →3 CasesCiri_aiView creator →3 CasesGoodmanprotocolView creator →3 CasesJust_sharon7View creator →3 Casesoggii_0View creator →3 CasestechhallaView creator →

A Skill is a derived reading layer. Every claim must trace back to published Cases.

Loading…

Good for

Inputs to prepare

—A specific place and time of day
—A description of the person, or a reference image to lock face and outfit
—Three to six small events in order
—One to three lines you want spoken
—Who holds the camera, her or a friend or partner
—Aspect ratio and length

Method steps

01Decide who holds the camera. State whose hand the phone is in and how far from the face, such as an arm's-length front camera, or a boyfriend holding it a metre or so away and never appearing.
02Split the length into segments and write what happens in each by the second. The Tokyo food, gym and cycling cases all set a timeline first, then the actions.
03Write in phone-camera flaws: walking bounce, slight shake, autofocus hunting, exposure shifts between inside and outside, the occasional cropped face.
04Put the exact spoken lines in the prompt and say they are delivered to camera with matching lip movement. An opening line and a goodbye are the most common.
05Ask for diegetic sound only, list concrete ambient sounds, and say no background music and no subtitles.
06End with bans such as no cinematic moves, no drone, no slow motion, no third-person camera, no beauty filter, and repeat that face and outfit stay the same throughout.

Variables to swap

—Setting and activity: festival, side-street dessert cafe, gym, train, mountain ride
—Who films: herself at arm's length, or a friend or partner filming her
—Device: phone front camera, rear camera, action cam, handlebar mount
—Lines: an opener, a mid-clip aside, a goodbye
—Aspect and length: 9:16 vertical or 16:9, 10 to 30 seconds
—One continuous take or a multi-shot mix

Common failures and limits

—Drifting into a cinematic look. Several creators add bans for exactly this: no cinematic moves, no drone shots, no color grading, no tripod or third-person camera. (Source: prompt avoid lists)
—Face or outfit changing between shots. Most prompts lock the same face and clothes again and again; the cycling prompt also rules out a third hand. (Source: prompt constraints)
—Lines may not come out as written. In our retest of the Tokyo food case the transcript has the scripted closing line, plus an opening line the prompt never wrote. The festival retest has no transcript, so we cannot tell whether the goodbye was spoken. (Source: retest data, single generation)
—Overdoing the flaws. The late-night kitchen prompt bans excessive shake; the boyfriend POV prompt bans digital zoom and face-filling close-ups. Keep shake and focus misses light. (Source: prompt avoid lists)
—The eight-shot cycling vlog, which switches between several camera types, scored 74.4 in our single retest; the Tokyo food and festival vlogs, which stay on one phone selfie view, landed around 82. One run each, so this says nothing firm about the cause. (Source: retest data, weak evidence)