AI Witch VideoYou in the '60s Witchcore Reel

Add one photo and you play the witch from the viral '60s witchcore reel: the rotary phone hang-up, smoking cauldron and black cat, with the original audio. Your cat or dog can join you, or play the witch itself.

The AI witch video is a fully AI-made '60s witchcore skit: a witch in a tall black hat hangs up a cream rotary phone, stirs a smoking cauldron and twirls as a black cat floats down beside her. Add one photo, and you play the witch, with the original audio.

Also called '60s witchcore AI witch · I have plans today witch · Witch on the rotary phone · Witch candle circle · Bestie tarot night

Example

What the AI Witch Video video looks like

This is the original reel. Your photos replace its characters, so your video follows the same shots with your faces. No two runs come out the same.

Original reel by @aiwitch1111, 9 s
The original AI reel by @aiwitch1111 (4 July 2026): the complete 9-second source begins with the rotary phone hang-up, then the witch stirs the cauldron and twirls as a black cat floats down. Your video follows that same sequence with the original audio.
Use this reference
The trend's template clip, 10 s
The trend's template clip for Candle circle: a witch kneels in a ring of candles in a dark forest, lights the last one, holds a thin blue candle to the lens, then the camera turns into the empty trees. Shown muted.
Use this reference
The trend's template clip, 10 s
The trend's template clip for Tarot duo: two friends in witch hats lay a card in a ring of tarot cards at dusk, blow out one candle together and pick a card. Shown muted.
Use this reference
Step by step

How to make the AI Witch Video video on Smew AI

Everything runs in the browser. No editing app, no prompt to write, no watermark.

  1. Add your photo

    Press "Create your own" and add your photo, face close and clear. Take it with the camera or pick it from your library.

  2. Press Generate

    The button shows the price. Seedance 2.5 puts your photos into the AI witch video trend clip and copies it shot by shot for 9 seconds, with the reel's original song. You write no prompt.

  3. Download or share

    The video is ready in up to several minutes as a 9:16 MP4. Download it, share it from your phone, or find it later in My videos.

Which photo works best?

Upload one photo for You or Pet, and two photos for You + pet (photo 1 is you, photo 2 is your pet).

  • You: chest up, face to the camera, sharp and in good light. The reel has two big close-ups of the face, so a close, clear photo matters most.
  • No sunglasses, no hat or hood that hides the hair line, and no beauty filter.
  • Your clothes do not matter: the witch hat, the long black dress and the turquoise winged liner come from the reel.
  • Pet: one animal, head and chest in view, eyes open, in daylight. The fur colors, ears and markings come from your photo.
  • Use your own photo, a friend who said yes, or your own pet. Never a celebrity or a stranger.

Where did the AI Witch Video trend start?

The reel comes from @aiwitch1111, an AI video creator on Instagram, who posted it on 4 July 2026 with the caption "They don't need to know what you're actually doing at home! This video is fully AI-generated! I'm absolutely in love with the '60s-style witchcore aesthetic." By 29 October it had 25,360 likes and 90 comments.

The joke is the contrast between an ordinary rotary-phone hang-up and the witch's real plans: a smoking cauldron and a floating black cat. The candle-lit red parlour, long black dress, sheer sleeves and turquoise winged liner create the 1960s witchcore look.

By late October 2026, AI template sites offered the skit as a one-photo template. This page copies the complete hang-up, cauldron, close-up and floating-cat sequence with the original audio.

  1. @aiwitch1111 posts the fully AI-generated '60s witchcore reel on Instagram: the rotary phone, "I have plans today", the cauldron and the floating black cat. Instagram

  2. Two days before Halloween, AI template sites offer the same skit as a one-photo template: you as the 1960s witch.

  3. The reel has 25,360 likes and 90 comments on Instagram. Instagram

  4. The #witchcore hashtag holds 58,538 TikTok videos and 439.6 million views, and #witchaesthetic 84,707 videos and 624.7 million views. TikTok

Copy and paste

The AI Witch Video prompts

You do not need these: the one-tap flow runs them for you, together with the trend clip. [Image 1] and [Image 2] are your two photos, and [Video 1] is the trend clip the video copies.

You

1 photo: your photo, face close and clearSeedance 2.59 s · 480P · 9:16, like the original183 credits
Video prompt

Use the complete [Video 1] as the performance reference. Preserve its actual duration, full shot sequence, framing, cuts, actions, expressions, mouth movements and camera timing; do not invent a new storyboard or retime the source. Replace only the witch with the person in [Image 1]: preserve that person's face, hair, skin tone, age and identity in every shot, including close-ups and the final twirl. Keep the reference witch costume, set, rotary phone, cauldron, black cat, lighting and all non-target subjects unchanged. Do not add dialogue or a different song: the generated picture is silent and the matching source soundtrack is restored server-side. Exactly one person and the original black cat, with natural anatomy and no text, captions, logos or watermark.

769 characters

Дуэт

2 photos: слушатель в наушниках (остаётся) + говорящий, который уходитSeedance 2.59 s · 480P · 9:16, like the original183 credits
Video prompt

Use the complete [Video 1] as the performance reference. Preserve its actual duration, full shot sequence, framing, cuts, actions, expressions, mouth movements and camera timing; do not invent a new storyboard or retime the source. Replace the witch with the person in [Image 1], preserving that person's face, hair, skin tone, age and identity in every shot. Replace only the floating black cat in the final shot with the pet in [Image 2], preserving its species, breed, face, ears, fur colors, markings and natural anatomy. Keep the reference witch costume, set, rotary phone, cauldron, lighting and camera unchanged. Do not add dialogue or a different song: the generated picture is silent and the matching source soundtrack is restored server-side. Exactly one person and one animal, no text, captions, logos or watermark.

826 characters

Вы + питомец

1 photo: ваш питомец в наушниках (остаётся)Seedance 2.59 s · 480P · 9:16, like the original183 credits
Video prompt

Use the complete [Video 1] as the performance reference. Preserve its actual duration, full shot sequence, framing, cuts, actions, camera movement and timing; do not invent a new storyboard or retime the source. Replace only the witch with the pet in [Image 1], preserving its species, breed, face, ears, fur colors, markings and natural body proportions in every shot; it remains an animal with paws, never a person or human hands. Fit the reference witch hat and black costume to the pet's body. Keep the rotary phone, cauldron, original floating black cat, set, lighting and camera unchanged. Do not add dialogue or a different song: the generated picture is silent and the matching source soundtrack is restored server-side. Exactly the uploaded pet and the original black cat, no people, text, captions, logos or watermark.

828 characters

Candle circle

1 photo: your photo, face close and clearWan 3.010 s · 480P · 9:1656 credits
1. Prepare the image · Nano Banana 2.1

Use the uploaded image as the base: the same person, now in a candle-lit witch ritual in a forest at night. An ordinary, unedited smartphone photo taken at night by a friend standing a few steps away, the phone at chest height, lit only by the warm light of candles. The place: a dark pine forest at night, dry pine needles and bare earth on the ground, tall straight tree trunks fading into black behind. The person kneels upright on a large black cloth in the middle of a circle of about twenty lit white pillar candles of different heights, with a few tarot cards laid face up on the cloth and a few red roses between the candles. The person holds one thin, lit, pale blue taper candle in both hands at chest height and looks straight into the camera with a calm, knowing, slightly mysterious look. Outfit: a tall black pointed witch hat with a wide brim and a long black lace dress with long sheer lace sleeves that covers the knees. A man wears the same hat and a long black robe with the same sheer lace sleeves. The kneeling person from the knees up in the middle of the frame, the nearest candles along the bottom edge, the face clear, sharp and lit warm by the candle flames, turned to the camera. Photorealistic, natural skin, real candle light and deep natural shadows: no cinematic color grading, no soft dreamy glow, no blurred bokeh, not CGI, not a cartoon. The tarot cards show plain old-style art with no readable text. Vertical 9:16. Exactly one person, no one else in the forest. No signs, no text, no logos, no watermarks.

This image is prepared automatically before the video step. Both steps are included in the displayed credit cost.

2. Animate the prepared image

An ordinary, unedited phone video filmed at night by a friend, in one continuous shot with no cuts, lit only by the candles, with a slight handheld shake, like a real clip on social media. Keep the person's face, hair, witch hat and black lace outfit exactly as in the image. 0 to 3 s: the person tips the thin blue candle down and lights the last unlit candle in front of them, then lifts the blue candle back to the chest. 3 to 6 s: the camera moves slowly closer as the person holds the candle still, the flame lighting the face from below, and looks straight into the lens with a calm, knowing look and a small slow smile. 6 to 10 s: the camera turns slowly away to the left, off the person, into the dark forest, and holds on the empty black trees and the bare ground, with only a faint warm candle glow on the nearest trunks: nobody is there. Audio: the natural sound of a quiet forest at night, the soft crackle of candle flames, a low cold wind in the pine trees and one distant owl. No music, no voices, no talking. One person only, five fingers on each hand, the candle flames stay small and calm, no fire spreads. No slow motion, no cinematic look, no text, no subtitles, no logos.

1192 characters

Tarot duo

2 photos: you (kneels on the left), face clear + your bestie (kneels on the right), face clearWan 3.010 s · 480P · 9:1656 credits
1. Prepare the image · Nano Banana 2.1

Use the two uploaded images as the base. The person in the first image kneels on the LEFT, and the person in the second image kneels on the RIGHT: two best friends at a witchy tarot night. Keep each one's own face, hair, skin tone and age; never blend or swap them. An ordinary, unedited smartphone photo taken at dusk by a friend a few steps away, the phone at chest height. The place: a garden lawn at dusk under a deep blue evening sky, dark hedges and trees behind. On the grass in front of them lies a big ring of about forty tarot cards laid face up edge to edge, a few more cards in the middle of the ring, tall white pillar candles at both sides and small warm tea lights dotted around the ring, all lit. The two friends kneel side by side, close together, behind the ring. The friend on the left holds one lit white pillar candle on a small dish between them at chest height; the friend on the right holds a small stack of tarot cards. Both look into the camera with a calm, playful, knowing smile. Outfits: the friend on the left wears a long deep-red ruffled dress with long sleeves, the friend on the right a long black lace dress with long sheer lace sleeves, and both wear tall black pointed witch hats; a man wears a dark shirt and a long black coat in place of the dress. Both from the knees up, the front of the card ring along the bottom of the frame, both faces clear and sharp, turned to the camera. Photorealistic, natural skin, real dusk light and candle light: no cinematic color grading, no soft dreamy glow, not CGI, not a cartoon. The tarot cards show plain old-style art with no readable text. Vertical 9:16. Exactly two people. No signs, no text, no logos, no watermarks.

This image is prepared automatically before the video step. Both steps are included in the displayed credit cost.

2. Animate the prepared image

An ordinary, unedited phone video filmed at dusk by a friend, in one continuous shot with no cuts, with a slight handheld shake, like a real clip on social media. Keep both friends' faces, hair, witch hats and outfits exactly as in the image: the one on the left stays on the left, the one on the right stays on the right. 0 to 3 s: the friend on the right leans forward and lays one more tarot card into the ring. 3 to 6 s: the friend on the left lifts the lit candle up between their two faces; both lean in close and blow it out together, and the flame goes out with a thin curl of white smoke rising between them. 6 to 10 s: they smile at each other, the friend on the right picks up one card from the ring and both look at it, eyebrows raised, then look into the camera with a knowing smile. Audio: the natural sound of a quiet garden at dusk, soft crickets, the soft blow of the two breaths and one short shared laugh. No music, no spoken words. Exactly two people, five fingers on each hand, both faces never morph. No slow motion, no cinematic look, no text, no subtitles, no logos.

1090 characters

How many credits does it cost?

You pay in credits. The Generate button shows this price before you spend anything. These are the options on this page.

OptionModelWhat you getCredits
YouSeedance 2.59 s · 480P · 9:16, like the original183
ДуэтSeedance 2.59 s · 480P · 9:16, like the original183
Вы + питомецSeedance 2.59 s · 480P · 9:16, like the original183
Candle circle: StandardWan 3.010 s · 480P · 9:1656
Candle circle: BestSeedance 2.510 s · 480P · 9:16128
Tarot duo: StandardWan 3.010 s · 480P · 9:1656
Tarot duo: BestSeedance 2.510 s · 480P · 9:16128

One-time packs and plans are on the pricing page. A failed video gives its credits back.

What happens in the AI witch video, shot by shot?

Your video follows the complete reference: the witch's hand hangs up the cream rotary phone, she stirs the smoking cauldron, appears in close-up, and twirls as the black cat floats down beside her. The cuts and timing stay with the source.

How it works: top-quality AI video models, including Seedance 2.5, edit the reel itself: they swap the witch for the person or pet in your photo and keep the set, costume, moves and camera. You get a 9-second vertical video at 480P, with the reel's original audio put back after the run. The reference begins with the hang-up; this template does not add an earlier phone conversation or new dialogue. A run takes 5 to 15 minutes, and the exact price shows on the Generate button before you spend anything. Every run is a new take, so results differ.

You, You + pet or Pet: which AI witch video should you make?

  • You puts your face on the witch in every shot. The costume is the joke, so it stays: the tall black pointed hat, the long black dress with sheer sleeves and the bold turquoise winged eyeliner. The prompt gives a man the same hat and a long black robe. This is how to turn yourself into a witch for Halloween without a costume.
  • You + pet makes you the witch and your own cat or dog the familiar: in the last shot your pet floats down beside you instead of the black cat. Photo 1 is you, photo 2 is your pet.
  • Pet makes your pet the witch in every shot: in the tall hat on the rotary phone, at the cauldron with the wooden spoon and in the twirl, while the black cat floats down beside it. A serious-looking dog or cat on a vintage phone is the funniest version.
  • Candle circle and Tarot duo are new witchy scenes, not the reel, and they work all year: you kneel in a ring of lit candles in a dark forest at night until the camera turns into the empty trees, or you and a friend (photo 1 left, photo 2 right) blow out one candle together behind a ring of tarot cards at dusk. Each is 10 seconds with the scene's own sounds, not the reel audio.

Caption ideas: "They don't need to know what I'm actually doing at home", "Me when someone asks if I'm around on Friday", "My plans for Halloween", "POV: you said you were busy". Or write your own.

More witch trends: fly past a car window in the My Ride Is Here broom trend, walk in the Ramalama Bang Bang witch trend, or try the dark enchantress filter. Just want a costume photo? Try the AI Halloween costume.

What goes wrong, and how do you fix it?

ProblemFix
You get long red hair, or the face looks like the woman in the reelThe witch in the reel has long copper-red hair, and the prompt bans it unless your photo shows it. A photo with red hair is the hardest case: use a sharp, close, front-facing photo and run it again.
The witch wears your own clothesThe prompt keeps the reel's witch costume. Run it again; a photo in plain clothes helps.
The mouth moves a little off the wordsThe mouth follows the reel, and the audio goes on after the run. Run it again for a new take.
A paw looks like a hand (Pet)The prompt asks for animal paws. Run it again with a photo where the whole head is clear.
The face is lost in the short shotsThe "Bye" shot lasts less than half a second. Check the close-ups; if one is off, run it again.
The model refuses a photoSome photos get refused. The credits come back; try another photo.

What is the witchcore aesthetic, and how should you post it?

Witchcore is the cozy, dark side of fashion and decor: black dresses, pointed hats, candles, old books, herbs in jars and black cats. This reel adds a 60s witch aesthetic on top: the soft colors and warm light of a 1960s TV show, a cream rotary phone, 60s witch makeup with turquoise winged liner and a 60s witch dress with flowing sleeves. If you look for 60s witch costume ideas or witchcore makeup and witchcore outfits for Halloween, the reel shows a full look you can copy.

Your clip comes with the reel's original audio, the same 9 seconds, so it is ready to post. You can turn it off on the result screen. The audio belongs to its owner, and Smew AI gives you no rights to it.

Hashtags: #aiwitch (1,834 TikTok videos and 10.6 million views on 29 October 2026), #witchcore (58,538 videos), #witchaesthetic (84,707 videos) and #halloween2026. Keep the platform's AI label on.

FAQ

AI Witch Video trend questions

What is the AI witch video trend?
It is a fully AI-generated '60s witchcore skit by the Instagram creator @aiwitch1111, posted on 4 July 2026. A witch hangs up a rotary phone, stirs a smoking cauldron and twirls as a black cat floats down. It had 25,360 likes by 29 October 2026, and people now make their own witch AI video from one photo.
Can AI turn me into a witch?
Yes. Pick You and add one clear photo. The AI turns you into the witch of the reel in every shot: your face and hair, the reel's witch hat, black dress and turquoise liner, the same phone hang-up, cauldron and twirl, with the original audio. You write no prompt.
Is there a turn me into a witch app?
Smew AI works in the browser on your phone or computer, so there is no app to install. It works like an AI witch app or an AI witch generator for video: add a photo, press Generate, and the video waits in My videos when it is ready.
Can my cat be the black cat in the AI witch video?
Yes. Pick You + pet: photo 1 is you as the witch, photo 2 is your cat or dog, which floats down beside you in the last shot instead of the black cat. That is the AI witch cat version. Pick Pet if you want your pet to play the witch itself.
Does the reference include the phone conversation?
No. The complete reference used here begins as the witch hangs up the rotary phone. It keeps that source's original audio and does not add the earlier conversation or any new spoken line.
Is the AI witch meme a real person?
No. The creator says the reel is fully AI-generated: the witch is an AI-made woman. In your video, your photo takes her place in every shot, and she does not appear.
How long is the video and what format is it?
9 seconds as an MP4 at 480P, vertical 9:16 like the reference, with the reel's original audio. It plays on TikTok, Reels and Shorts, and there is no watermark.
Which prompt does this template use?
The creator did not share a prompt, and the template sites that copy the skit hide theirs. This template uses its own prompt: it copies the reel shot by shot, swaps only the witch, keeps the costume and adds rules that keep your face from the photo, or your pet's fur, ears and markings.
Does the AI witch video cost credits?
Yes. Each video costs credits, and the create page shows the exact price on the Generate button before you spend anything. If a video fails, the credits come back.
Is there a Halloween witch video with a witch on the phone and a witch with black cat in one?
Yes, that is this one: the witch on the phone opens the reel and the witch with black cat closes it. It is a short AI Halloween video made from the reel, with your face in it.
Is there a witch photoshoot with friends or a witch candle ritual version?
Yes, two versions that work all year, not only at Halloween. Candle circle is a forest witch photoshoot at night: you kneel inside a witch ritual circle of lit candles, hold a thin candle and look into the lens, then the camera turns away into the dark, empty trees, a witch candle ritual in one photo. Tarot duo is a witch photoshoot with friends and a tarot photoshoot in one witchy photoshoot: you and your bestie kneel behind a ring of tarot cards at dusk and blow out one candle together. Photo 1 kneels on the left, photo 2 on the right. Each makes a 10-second video with the scene's own sounds and no music. AI template apps started to offer both scenes on 30 October 2026.
Can a man make the AI witch video?
Yes. The video keeps the gender of your photo, so a man gets the same tall witch hat and a long black robe with the same sleeves. Some people call the result an AI witch girl video, but anyone can play the witch.

Make your own AI Witch Video video

Add one photo and tap once. No prompt to write. 183 credits.

No watermark. A failed video gives its credits back. Secure checkout by Stripe.