Turn your image into a Ramalama Bang Bang witch AI video with a reveal worth waiting for. A witch flies across the full moon, a phone captures the silhouette, and a camera flash reveals you riding the broom. Use copy-ready prompts to create your own moonlit scene, try a cat or dog as the witch, or follow a reference clip in a compatible video workflow. Finish with playful broom poses for TikTok, Reels, or Shorts.
Start with one clear person or pet image. Use the prompts below for a new scene, or add a reference clip in a mode that supports it. Add the Ramalama soundtrack and align the reveal during editing.
Make the Ramalama witch trend personal: keep the rider hidden in silhouette until the flash reveals your face. Use your image to guide facial features, skin tone, and hairstyle, or your pet’s distinctive eyes, ears, and fur markings.
Build the taking a picture witch trend around its key sequence: a broom crosses the moon, hands raise a phone, and a flash cuts to the rider. Save the recognizable close-up for the reveal so the surprise has room to land.
Give your witch broom AI video a clear visual anchor: one rider on one broom against a huge full moon. Request flowing fabric, gentle hovering, and a stable seated pose, with the face visible beneath the witch hat.
Create a 9:16 Halloween AI video with space for the moon, phone, and broom. Keep the opening silhouette readable on a small screen, then bring the face closer for the final poses without cropping the hat or broom.
Upload one clear, well-lit image with the face unobstructed. A front-facing portrait helps define the rider’s identity; a full-body image can also guide proportions. For a pet version, choose one cat or dog with its face, ears, and markings clearly visible.
Choose a prompt below to plan the full moon, phone flash, and witch reveal. Use Image 1 for identity. For closer control over an existing clip, add Video 1 only in a compatible reference-video mode; otherwise, generate a new interpretation from your image and prompt.
Preview your Ramalama Bang Bang AI video for face consistency, phone shape, and the rider’s position on the broom. Download your favorite take, add the selected Ramalama sound, and trim the flash and final poses to the music before sharing.
One portrait + a moonlit witch reveal
Go from a mysterious silhouette to the unmistakable face on the broom. Prompt: “Create a vertical 9:16 Halloween video using Image 1 as the identity reference for one adult witch. Preserve the person’s face, skin tone, hairstyle, and proportions. Opening: a distant witch on a wooden broom flies sideways across a huge full moon above a quiet city. Keep the face hidden in silhouette. Next, two hands enter the foreground holding a smartphone pointed at the moon; only the photographer’s hands are visible. A brief white camera flash transitions to a medium shot of the same witch seated sideways on the hovering broom in front of the moon. Reveal the recognizable face from Image 1, clearly lit beneath a black pointed hat. The witch smiles, tips the hat, and settles into a playful pose. Keep the black outfit and cape consistent, with gentle fabric movement. One witch, one broom, one moon. SAME IDENTITY. NO DUPLICATE. No extra scenes, face morphing, extra limbs, body-broom fusion, text, or logos.”
One portrait + a reference-video workflow
Follow your chosen Ramalama clip’s camera, flash, and broom poses. Prompt: “Use Image 1 as the sole identity reference and Video 1 as the Ramalama witch scene and motion reference. Replace only the witch’s identity with the person in Image 1. Keep their facial structure, skin tone, eyes, and hairstyle recognizable. Follow Video 1 for the moon-crossing silhouette, foreground phone, camera flash, close-up reveal, seated broom poses, shot order, camera movement, and timing. Keep the reference witch costume, broom, moon, background, and lighting. Do not replace the photographer’s hands with the rider’s hands. Keep the witch’s face hidden until the reference reveal, then maintain the same identity in every visible close-up. Match the reference aspect ratio and duration within the selected mode’s limits. SAME IDENTITY. NO DUPLICATE. NO NEW SCENES. No original rider’s face, extra witches, extra brooms, distorted hands, body-broom fusion, costume flicker, added text, or logos.”
One cat or dog image + a broom reveal
Let the flash reveal your pet as the tiny witch in the moonlight. Prompt: “Create a vertical 9:16 Halloween witch reveal using the single pet in Image 1. Preserve its species, breed, face, eyes, muzzle, ears, coat colors, and fur markings. Start with a small witch silhouette riding a broom across a large full moon. Foreground human hands raise a smartphone toward the moon; show no photographer’s face or body. A camera flash cuts to the reveal: the pet from Image 1 sits securely on the floating broom, wearing a pet-sized black witch hat and cape. Its face and ears remain visible. The pet turns its head toward the camera, blinks, and tilts its head playfully while the broom gently hovers. Keep its natural animal proportions, four limbs, and tail where present; no human hands or human face on the pet. Keep the moonlit setting and costume consistent. One pet rider and one broom. SAME PET IDENTITY. NO DUPLICATE. No extra scenes, changing fur, extra paws, animal-broom fusion, text, or logos.”
One portrait + a shorter reveal sequence
Keep the suspense, flash, and final pose in a compact vertical edit. Prompt: “Create a short vertical 9:16 Ramalama-style witch reveal with the adult in Image 1 as the only rider. Preserve their exact facial features, skin tone, and hairstyle. Use three clear beats: first, a witch silhouette on a broom crosses the full moon; second, a pair of foreground hands holds up a phone and takes a picture with a brief flash; third, cut to the rider’s recognizable face as they sit sideways on the broom in front of the moon. End with a confident smile and one small hat-tip, leaving a brief hold for editing to music. Keep the same black witch hat, flowing cape, wooden broom, and moonlit sky throughout. The photographer stays off-screen except for the hands. One witch and one broom. SAME IDENTITY. NO DUPLICATE. No extra scenes, costume walk, dancing on the ground, face morphing, extra limbs, body-broom fusion, text, or logos.”
The Ramalama Bang Bang witch AI trend turns a person or pet into the rider on a flying broom. A silhouette crosses the full moon, a phone takes a picture, and a flash reveals who the witch is before a few playful poses. It uses the same song as the Ramalama costume walk, but the action here stays in the moonlit sky rather than following a person’s feet along the ground.
Choose one sharp image with the face clearly visible and minimal obstruction from hats, sunglasses, or filters. A portrait works for a person; a full-body image can provide more information about proportions. For a cat or dog, use an image that clearly shows the face, ears, and coat markings. The pet prompt adapts the witch costume and seated broom pose to the animal. Use an image you have permission to animate, and check the generated anatomy before sharing.
You can use an image and a scene prompt to generate a new version of the moonlit witch reveal. To follow a specific clip’s camera movement, flash timing, and poses more closely, use a mode that accepts both an identity image and a reference video. In that workflow, Image 1 defines the rider and Video 1 guides the sequence. A prompt containing the trend name alone does not supply the original choreography or guarantee the same result.
Use 9:16 for a vertical TikTok, Reel, or Short, leaving enough room for the moon, raised phone, and broom. Choose a duration supported by your selected model and settings. Keep the opening silhouette brief, give the flash a clear transition point, and leave time to recognize the face after the reveal. If a single generation struggles with the full sequence, generate the opening and reveal separately and join them with a flash transition.
The trend uses “Ramalama (Bang Bang)” by Róisín Murphy. Do not assume your generated video includes the original recording or automatic beat matching. Add an available, authorized version of the sound in your editing or social app, then align the phone flash, face reveal, and final poses with your selected excerpt. These prompts describe the visuals and do not include a soundtrack or music license.
Create Images and Videos
