wowprompt
A young woman in a black hoodie looking up at a corner security camera in a cool-lit corridor at night

CCTV AI Photo Prompts: 6 Security Camera Looks (2026)

CCTV AI photo prompt guide for ChatGPT and Gemini: high corner angle, fisheye bend, timestamp overlay, night vision and six copy-ready security camera looks.

A CCTV AI photo prompt turns your own photo into a picture that looks like a still from a security camera: shot from a high corner or through a fisheye doorbell lens, stamped with a camera label and a running timestamp, and graded in the cool fluorescent or green night-vision cast of real surveillance footage. The joke works because the moment is ordinary and the framing is serious: a bride laughing in a lift, a 2 AM noodle run down a hostel corridor, caught on camera as if by accident.

Below: what makes the look believable, the mistakes that give it away, how to use it with your own photo, and six ready prompts with sample pictures.

What the CCTV photo trend is

The trend recreates a frame grabbed from surveillance footage: a candid, slightly distant, slightly distorted view of you doing something everyday, with on-screen text that says where and when. Three things define it:

  • The camera position. High in a ceiling corner looking down, or at chest height in a door looking out.
  • The overlay. A small camera label such as "CAM 03", a date and a running time with seconds.
  • The grade. Flat, cool colour under tube lights, or green and grey when infrared night mode is on.

It sits close to candid portrait prompts and the film still look, like our 1990s film still in a velvet bandhgala down a hotel corridor, but trades cinematic light for the honest light of a building at night.

What makes a CCTV AI photo prompt look real

Most failed attempts look like a normal photo with a timestamp pasted on top. A believable still gets six details right.

Angle and height

Put the camera where real cameras live. A lift camera sits in the top corner opposite the door. A corridor camera hangs from the ceiling at one end and looks down the length of the hall. A shop camera sits above the counter. A doorbell camera sits at chest height and looks slightly up at the visitor. Write the position into the prompt in plain words: "seen from a ceiling camera in the top left corner, looking down at about 45 degrees".

Lens bend

Security cameras use wide lenses, so straight lines bow near the edges and people near the frame edge stretch slightly. Ask for "a wide fisheye lens with barrel distortion at the edges" for a lift or a doorbell, and a milder "wide-angle lens with slight edge bend" for a corridor or a shop. The fisheye selfie with prayer flags on a Ladakh pass shows how strongly a fisheye curves a scene.

Overlay type

Keep the on-screen text short and generic: a camera name in letters and digits ("CAM 04", "FRONT DOOR"), a date and a time, in white blocky text in one corner. A small red "REC" dot is optional.

Timestamp format

Spell out one format, because models invent odd dates when left alone. "2026-11-08 02:14:37" reads as a real recorder, and digits come out cleaner than month names. The time also tells the story: 02:14 says hostel, 21:42 says wedding reception.

Grade and grain

Real CCTV colour is flat and slightly cool, with clipped highlights on tube lights. Ask for "low dynamic range, mild compression blocks, soft digital noise". In night mode, ask for "infrared night vision, monochrome green and grey, deep black corners".

The face still clear

This is where the trend fails most often. A real camera would make your face small and soft; your picture should not. Ask for "the face clear and recognisable despite the camera angle" and keep the subject in the near half of the frame, turned partly toward the lens. A laugh, a glance up or a look back gives the model a reason to show the face.

Common mistakes with the security camera AI photo

  • Night vision everywhere. Green infrared only makes sense in the dark, such as a basement car park or an unlit gate. A lift or a shop under tube lights records in colour.
  • Unreadable faces. A tiny figure at the far end of a corridor is realistic and useless. Bring the subject closer and have them look up or back.
  • Fake brand overlays. Leave out camera maker names, logos and shop names. Invented brand text is where spelling errors appear. A generic label is enough.
  • Orange cast. Warm filters make the still look like a phone photo. Put "orange cast" in the avoid line of every prompt.
  • Too clean or too wordy. A perfectly sharp frame reads as a poster, and a long overlay turns into gibberish. Ask for mild noise and keep the text to a label, a date and a time.

How to use a CCTV photo trend prompt with your own photo

  1. Open ChatGPT or Gemini and start a new chat.
  2. Attach one clear, front-facing photo in even light. For the couple still, attach a photo with both people.
  3. Paste a prompt from below as written, keeping the identity line and the avoid line.
  4. Check the face first. If it drifted, reply: "Keep everything, but make the face match my photo exactly."
  5. If the timestamp came out wrong, reply: "Keep the image; correct the timestamp to read 2026-11-08 02:14:37."

ChatGPT tends to follow the overlay and camera position more closely; Gemini often holds the face closer. Our reference photo guide explains why the identity line comes first.

Six CCTV AI photo prompts to copy

Each section has the full prompt and the picture made from it.

1. Bride and groom laughing in a hotel lift

A fisheye view from the top corner of a hotel lift, in cool fluorescent light, as a bride and groom in reception clothes share a private laugh on the way up. Use it for a wedding reel or an anniversary post.

With a photo attached: the two people in the photo are the subjects. Keep each face exactly as it is, with bone structure, eye shape and spacing, nose, lips, jawline, hairline, skin tone and apparent age all unchanged, and keep each person's own body: build, height and proportions, shoulders, arms, hands and legs, with the same skin tone on face, neck, arms and hands. Use the photo only for who they are. Ignore its poses, expressions, clothing, hairstyles, background, framing and light, and rebuild all of them exactly as described below, even when the photo shows them standing straight or dressed differently. If the photo shows only one person, keep that face and create the other as described below.
Logline: Wedding night, the ride up from the reception. A bride and groom alone in a hotel lift, laughing at a shared secret, caught by the lift's security camera. The key idea: it reads at a glance as real CCTV footage, high corner, fish-eye, cool and plain, yet the couple look their best and both faces are clear.
Subject: They lean against the back wall. She tilts her head up toward the camera with a conspiratorial grin, right index finger raised to her lips, her left hand holding his lapel. He bends toward her ear, mid-laugh, eyes creased shut, his right hand holding his folded safa, his left arm round her waist.
Wardrobe: She wears a silver-grey tissue lehenga with tonal sequins, the dupatta bunched over her left arm. He wears an ivory raw-silk sherwani and a thin pearl mala.
Set: Brushed steel walls, a handrail, a lit floor-number panel by the door showing 7.
Text: - "LIFT 2 CAM 01" - "24-11-2026 23:47:12" Small white monospace type, top left, nothing else written.
Lens and grade: Fish-eye from the top corner, slight barrel distortion, the couple centred and least distorted. Cool fluorescent cast, soft video grain, faint scan lines, real skin texture.
Avoid: airbrushed skin, plastic skin, waxy face, changed face, distorted hands, extra fingers, green night vision, black and white, heavy pixelation, unreadable faces, orange cast, extra people, misspelled text, extra text, logo, watermark
Aspect ratio 4:5.
A high corner view of a bride in a silver-grey lehenga and a groom in an ivory sherwani laughing in a steel hotel lift, a timestamp in the corner
CCTV still of a bride and groom laughing in a hotel lift

Open this prompt on its own page: CCTV still of a bride and groom laughing in a hotel lift

For more lift moments, try the couple mirror selfie in a vintage brass cage elevator or browse the couple library.

2. Hostel corridor at 2 AM with instant noodles

A ceiling camera looks down a long hostel corridor as a girl in pyjamas carries a steaming cup of instant noodles back to her room at 2 AM. It suits college memories or a funny exam-week post.

With a photo attached: the person in the photo is the subject. Keep their face exactly as it is, with bone structure, eye shape and spacing, nose, lips, jawline, hairline, skin tone and apparent age all unchanged, and keep their own body: build, height and proportions, shoulders, arms, hands and legs, with the same skin tone on face, neck, arms and hands. Use the photo only for who the person is. Ignore its pose, expression, clothing, hairstyle, background, framing and light, and rebuild all of them exactly as described below, even when the photo shows them standing straight or dressed differently.
Scene: Two in the morning in a girls' hostel in Pune, second floor, B block. The key idea: it reads at a glance as a real corridor security camera frame, high on the ceiling, wide and plain, yet she has spotted the lens and grins up into it like an accomplice, her face sharp and flattering.
The moment: She walks toward the camera down the middle of the corridor, caught mid-step, torso upright, head tipped right back so her whole face looks up into the lens, eyebrows raised, a wide guilty grin showing her teeth. Her right hand holds a paper cup of instant noodles at chest height, steam curling up past her chin; her left hand pinches a plastic fork raised in a little salute. Bare feet in grey rubber slippers. She wears a heather-grey cotton fleece hoodie two sizes too big, sleeves pushed to the elbow, navy and white checked cotton pyjama bottoms, no jewellery, no makeup, a hair tie on her wrist. The corridor runs back behind her: cream-painted walls, numbered green doors, a steel water cooler, a drying towel on one door handle, a chappal stand.
Text: - "BLK-B FL2 CAM 03" - "18-11-2026 02:07:41" Small white monospace type, top left, nothing else written.
Look: Wide lens from the ceiling corner with mild barrel distortion, her face near the centre and least distorted. Cool fluorescent tube light, a faint green tint in the walls, soft video grain, light scan lines, real skin texture.
Avoid: airbrushed skin, plastic skin, waxy face, changed face, distorted hands, extra fingers, orange cast, warm wash, night vision, black and white, heavy pixelation, blurred face, extra people, readable brand on the cup, misspelled text, extra text, logo, watermark
Aspect ratio 4:5.
A high corner view of a young woman in a grey hoodie and checked pyjamas in a hostel corridor grinning up at the camera holding a steaming noodle cup
Hostel corridor CCTV still of a girl with 2 AM instant noodles

Open this prompt on its own page: Hostel corridor CCTV still of a girl with 2 AM instant noodles

The same corridor energy appears in the 18th birthday at midnight in a hostel corridor under phone torches.

3. Doorbell camera view of a woman in a red saree with sweets

An eye-level fisheye doorbell camera catches a woman in a deep red saree leaning in close with a box of sweets, as if she has just rung the bell. A doorbell camera AI photo suits a festival visit or a fashion post with a twist.

With a photo attached: the person in the photo is the subject. Keep their face exactly as it is, with bone structure, eye shape and spacing, nose, lips, jawline, hairline, skin tone and apparent age all unchanged, and keep their own body: build, height and proportions, shoulders, arms, hands and legs, with the same skin tone on face, neck, arms and hands. Use the photo only for who the person is. Ignore its pose, expression, clothing, hairstyle, background, framing and light, and rebuild all of them exactly as described below, even when the photo shows them standing straight or dressed differently.
Brief: A Saturday morning visit to a friend's flat in a Pune housing society. She rings the bell, then leans right up to the little doorbell camera to check whether it is recording. The key idea: a real doorbell camera frame, eye level and fisheye, so her face looms large and curious at the centre while the corridor curves away behind her.
Blocking: She stands a forearm's length from the lens, torso bent forward from the hips, head tilted slightly left, eyes wide and looking straight into the lens, one eyebrow raised, lips pressed in a small amused smile. Her right index finger presses the bell button at the lower edge of the frame. Her left hand lifts a white cardboard sweet box tied with red string beside her cheek.
Wardrobe: A deep red handloom cotton saree with a narrow black and white checked border, crisp pleats, the pallu pinned at the shoulder; a black cotton blouse with elbow sleeves and a round neck; oxidised silver jhumkas, a small black bindi, kajal, bare lips.
Set: A society corridor with a speckled grey mosaic floor, a cream wall, a neighbour's steel shoe rack and a potted tulsi in the background, a window at the far end.
Text: - "FRONT DOOR" - "2026-11-21 11:42:09" Small white monospace type, top left and bottom right, nothing else.
Camera: Wide fisheye at eye level with strong barrel curve and slight vignette, her face sharp. Cool clean daylight, true red, faint compression grain, real skin with pores.
Avoid: airbrushed skin, plastic skin, waxy face, changed face, distorted hands, extra fingers, orange cast, warm wash, festival decorations, diyas, rangoli, silk saree, heavy gold jewellery, warped face, extra people, misspelled text, extra text, logo, watermark
Aspect ratio 4:5.
A fisheye doorbell view of a young woman in a deep red cotton saree leaning toward the lens, ringing the bell and holding a box of sweets
Doorbell camera view of a woman in a red saree with sweets

Open this prompt on its own page: Doorbell camera view of a woman in a red saree with sweets

More saree and festive looks live in the editorial and fashion library and the traditional wear library.

4. A 3 AM laugh at a Bengaluru chai shop

A colour shop camera above the counter catches a young man in a hoodie laughing with the chaiwala at 3 AM. Use it for a friendship post or a late-night city story.

With a photo attached: the person in the photo is the subject. Keep their face exactly as it is, with bone structure, eye shape and spacing, nose, lips, jawline, hairline, skin tone and apparent age all unchanged, and keep their own body: build, height and proportions, shoulders, arms, hands and legs, with the same skin tone on face, neck, arms and hands. Use the photo only for who the person is. Ignore its pose, expression, clothing, hairstyle, background, framing and light, and rebuild all of them exactly as described below, even when the photo shows them standing straight or dressed differently. The subject is the customer; create the chaiwala as described below.
Base: A 24-hour chai shop on a quiet Bengaluru main road at three in the morning, seen from the camera bolted high in the corner behind the counter. The key idea: it reads as plain colour security footage, but it caught a real moment of friendship between a regular and the man who makes his tea.
Add: the customer: He leans on both forearms on the counter, torso angled toward the camera, head thrown back and turned up toward the lens, eyes creased, laughing with an open mouth. His right hand holds a small steel tumbler of tea; his left hand slaps the counter top. He wears a bottle-green cotton hoodie with the hood down, a white t-shirt under it and black track pants.
Add: the chaiwala: Smaller and softer at the right edge, in a faded blue checked shirt and a cream towel over his shoulder, pouring tea from a raised steel pan into a glass, grinning at him.
Add: the shop: A dented aluminium kettle on a gas burner, glass jars of rusk and biscuits, a stack of steel tumblers, an empty road outside the open shutter.
Text: - "CHAI CTR CAM 2" - "07-12-2026 03:12:58" Small white monospace type, top right, nothing else.
Grade: Colour CCTV look, high corner, wide lens, slight distortion at the edges. Cool blue-white tube light, the burner flame the only warm point, soft video noise, real skin texture, both faces clear.
Avoid: airbrushed skin, plastic skin, waxy face, changed face, distorted hands, extra fingers, orange cast, warm wash, night vision, black and white, green tint, blurred faces, readable shop signs, brand labels, extra customers, misspelled text, extra text, logo, watermark
Aspect ratio 4:5.
A high corner colour CCTV view of a young man in a hoodie laughing at a steel chai counter at night while an older chaiwala pours tea
Security camera still of a 3 AM laugh at a Bengaluru chai shop

Open this prompt on its own page: Security camera still of a 3 AM laugh at a Bengaluru chai shop

For a brighter take on the same city at night, see the phone flash night portrait on a scooter outside a Bengaluru darshini.

5. Night-vision CCTV of a woman rider in a Mumbai basement parking

Green infrared night vision in a dark basement car park, as a woman rider walks away from her motorcycle and glances back at the camera, helmet in hand. This is where night vision belongs, and it makes a strong rider profile picture.

With a photo attached: the person in the photo is the subject. Keep their face exactly as it is, with bone structure, eye shape and spacing, nose, lips, jawline, hairline, skin tone and apparent age all unchanged, and keep their own body: build, height and proportions, shoulders, arms, hands and legs, with the same skin tone on face, neck, arms and hands. Use the photo only for who the person is. Ignore its pose, expression, clothing, hairstyle, background, framing and light, and rebuild all of them exactly as described below, even when the photo shows them standing straight or dressed differently.
Must hold: Green infrared night-vision monochrome only, no other colour. Her face must stay clear and recognisable in the glow, turned three quarters back toward the camera. The motorcycle carries no badge, logo or readable plate. The key idea: the bright, flat infrared look of a real basement security camera, used to make her exit look like a film still.
The frame: Seen from a camera high on a concrete pillar, she walks away down the lane between parking bays, mid-stride, left foot forward. Her torso faces away, shoulders squared; her head turns back over her right shoulder, chin lowered, eyes straight into the lens, a faint knowing half-smile. Her right hand carries a full-face helmet by the chin bar at her hip; her left hand is pushing her hair back off her forehead. She wears a black textile riding jacket with armour panels and reflective piping, slim black jeans and black ankle riding boots. Behind her, parked at an angle in bay B2-14, a plain tank-and-round-headlamp motorcycle, engine still ticking. Painted bay numbers, a ramp sign arrow, low ceiling pipes, two parked cars soft in the dark.
Text: - "B2 PARKING CAM 14" - "IR ON" - "22-11-2026 23:58:31" Small white monospace type, top left, nothing else.
Look: Infrared glow on skin and reflective piping, eyes catching the light, gentle vignette, video grain, real skin texture.
Avoid: airbrushed skin, plastic skin, waxy face, changed face, distorted hands, extra fingers, full colour, orange cast, motorcycle logo, brand badge, readable number plate, blurred face, face hidden by helmet, extra people, misspelled text, extra text, logo, watermark
Aspect ratio 4:5.
A green night-vision view of a young woman in a riding jacket walking away from a motorcycle in a basement car park, looking back at the camera
Night-vision CCTV of a woman rider in a Mumbai basement parking

Open this prompt on its own page: Night-vision CCTV of a woman rider in a Mumbai basement parking

Riders will also like the motorcycle rider on a rain-wet street at night, helmet in hand and the cars and bikes library. For another car park mood, see the skipping rope workout in an empty concrete parking garage at night.

6. "I tried the CCTV photo trend" thumbnail

A YouTube thumbnail for creators covering the trend: a surprised face beside a night-vision still, built to read at a small size. Use it for a reaction video or a tutorial.

With a photo attached: the person in the photo is the subject in both parts. Keep their face exactly as it is, with bone structure, eye shape and spacing, nose, lips, jawline, hairline, skin tone and apparent age all unchanged, and keep their own body: build, height and proportions, shoulders, arms, hands and legs, with the same skin tone on face, neck, arms and hands. Use the photo only for who the person is. Ignore its pose, expression, clothing, hairstyle, background, framing and light, and rebuild all of them exactly as described below, even when the photo shows them standing straight or dressed differently.
Subject: A YouTube thumbnail split into two parts. The right half is the creator's face, close and bright. The left half is a tilted rectangular panel showing a green night-vision CCTV still of the same man from a high corner camera. The key idea: the reaction and the result side by side, the same person in both.
Text: - "I TRIED" - "THE CCTV PHOTO TREND" Nothing else is written outside the panel. Inside the panel, only a small overlay: "CAM 04  11-11-2026  02:14:37". Spell each line exactly.
Pose: In the right half he faces the lens, shoulders up, head pulled back a little, eyes wide, mouth open, both hands raised beside his cheeks with palms out. In the panel he stands in a narrow alley looking up at the camera, hands in his jacket pockets.
Wardrobe: A black nylon bomber jacket over a grey tee, a thin silver chain.
Setting: The face sits on a plain deep charcoal background. The panel shows a wet alley at night, pipes and a shut shutter, in grainy monochrome green with a slight fish-eye bend.
Light: A soft cool key on his face from the left, a thin cyan rim on the right.
Camera: An 85mm look at f/2.8 for the face, real skin with pores. "I TRIED" in heavy white condensed capitals above the panel, "THE CCTV PHOTO TREND" below it in black on a bright yellow band.
Avoid: misspelled text, extra words, logos, app icons, two different faces, text over face, clean sharp cctv frame, colour cctv frame, orange cast, golden glow, warm wash, airbrushed skin, plastic skin, waxy face, changed face, distorted hands, extra fingers, watermark
Aspect ratio 16:9.
A surprised young man in a black bomber jacket beside a green night-vision CCTV still of himself in a wet alley seen from above
I tried the CCTV photo trend thumbnail beside a night-vision still

Open this prompt on its own page: I tried the CCTV photo trend thumbnail beside a night-vision still

More layouts for video covers are in the thumbnail prompt library.

Ideas to vary the CCTV photo trend

Once one still works, change one element at a time:

  • Swap the place. A society gate, a gym floor, an office pantry, a metro platform. Every building has a camera.
  • Swap the time. 06:02 on a morning run, 23:58 just before a birthday, 04:30 after a night shift.
  • Make a sequence. Ask for "CAM 01", "CAM 02" and "CAM 03" views of one moment from three corners and post them as a carousel.
  • Split the screen. A four-way monitor grid, each view with its own label, works well for a group of friends.
  • Add a reflection. A lift mirror shows the face from the front and the camera angle from above, as in the mirror selfie in a dim lift with a film camera.
  • Turn it black and white. Older cameras record in grey. Read our black and white AI photo prompts guide for grading tips.

Frequently asked questions

What is a CCTV AI photo prompt?

It is a prompt that asks an AI model to turn your photo into a still that looks like security camera footage, with a high corner or fisheye angle, a camera label and timestamp overlay, and a cool fluorescent or green night-vision grade.

Does the CCTV photo trend prompt work in both ChatGPT and Gemini?

Yes. Paste the same prompt into either one with your photo attached. ChatGPT usually follows the overlay and camera position closely, and Gemini often keeps the face closer to your photo.

How do I keep my face clear in a security camera AI photo?

Keep the identity line first, place yourself in the near half of the frame, and give a reason to look toward the lens, such as a laugh or a look back.

When should I use night vision instead of colour?

Only when the scene is dark, such as a basement or an unlit gate. Lifts, corridors, shops and lit doorways record in colour.

Can I make a doorbell camera AI photo with a friend or family member?

Yes. Attach a photo with both people, ask for both faces to be kept, and place them side by side close to the door so the fisheye bend falls on the background.

Where to find more

For other trending formats, see our trending photo editing prompts, the nine-frame contact sheet AI photo prompts and the all-round ChatGPT photo editing prompts. Or browse every look in the image prompt library.

cctv ai photo promptcctv photo trend promptsecurity camera ai photodoorbell camera ai photonight vision ai photochatgpt cctv promptgemini cctv prompt

Read next