Tsumugi's🔨WishMallet Help

This app turns one photo and a short description into a video with a voice, entirely inside your own PC. This page explains the screens and how to create a video step by step, so you can follow it even if this is your first time.

🚀Your first video in 3 steps

Once setup is finished, try making one video with a sample photo. The tour that appears the first time you start the app guides you through the same steps on the screen.

①Add one photo“+ Add” on the “Create” tab → “🎁Sample files” → 01_tsumugi.jpg
②Write a few wordsA few words in “②Video description” → “🪄Auto prompt”
③Create the video“Create the video” → when it is ready, click its picture to play it

In ② you only write a few words about the video you want to see (for example A beautiful female samurai quietly readies her sword in a storm of cherry blossoms; the tour puts this example in the box for you). Detail, quality and the other settings can stay as they are (the length starts at 1 second, which finishes fastest). To build the description by answering questions instead, use 🧭Guided writing (→ ②in 3.). To start from an example, press 💡Examples under the ② heading and pick one: its description and settings (and the photo or voice) are filled in (→ Examples).

The detailed steps are in 3. How to create a video, and where everything is on the screen is in 2. The screens.

1What this app can do

How it works: a photo, a voice (optional) and a description go in, and a video with a voice is made inside your PC 🖼Photo 🗣Voice (optional) 📝Description 💻 Made insideyour PC 🎬 Videowith a voice
Your photos, voices and descriptions are never sent over the internet (with the built-in LLM, the default. If you format with OpenRouter, the description and similar text are sent over the internet)

To run the app you need Windows 11, an NVIDIA graphics card (12 GB or more of video memory), 64 GB or more of memory and 105 GB or more of free disk space. It does not run on AMD (Radeon) or Intel graphics cards. You can check before you buy with the System check (Web quick version). On a PC that doesn't meet these, or on a Mac or phone, you can use the web version in your browser (videos are made on a GPU rented on Runpod).

🎁Before you buy, you can use the 7-day free trial (up to 3 videos a day). See 8. The Settings screen, “Free trial”.

2The screens

The diagram below is a simplified view of the screen when the “Create” tab is open. The red circles match the explanations under the diagram.

Use the tabs along the top to switch screens.

TabWhat the screen is for
▶️CreateAdd a photo and a description, and create a video. The progress while it is being made and the finished video also appear here.
⏳QueueYour queued videos. Make them in order from the top, pause, or remove them (→ Making several videos in a row).
🎞️HistoryA list of the videos you have made so far. You can watch them, re-create them, cut out favorite moments and more.
🖼️AlbumA photo album (Best shots) of your “favorite single frames” cut out of your videos.

The following buttons appear at the right end of the tabs on every tab (except 🧹Clear all inputs, which is on the right of the ② heading on the “Create” screen; 💾Save to history is in the ☰Menu to the right of Create the video at the bottom of the “Create” screen).

ButtonWhat it does
🧹Clear all inputsClears the reference pictures, voices, reference videos, the description (including the original text that “Original” shows) and everything typed in Guided writing (including earlier entries), and puts detail, length, shape and quality back to their first values (you will be asked to confirm; the photo and video files on your PC are not deleted). To clear only what you typed in the guide, use the button with the same name inside the guide.
💡Handy tipsOpens “Handy tips” on this page.
❓HelpOpens this page. When the app language is English, this English page opens; when it is Japanese, the Japanese page opens (the same goes for “💡Handy tips”).
⚙️SettingsLanguage (Japanese / English), save location, license, diagnostics for when something goes wrong, and more (→ The Settings screen).

Hover the mouse over a button or an input box to see a short explanation. When in doubt, hover first.

3How to create a video

Open the “Create” tab and work from top to bottom.

The first time you start the app, a short walkthrough (the tour) appears over the screen. Press “Next” and the buttons it points to (+ Add, 🪄Auto prompt and Create the video), and it takes you all the way to your first video. If you do not need it, you can close it right away with the “×” at the top right, “End tour”, the Esc key or a click on the dark background (it will not appear again). To see it again, go to “Settings” → “About” → Take the tour.

①Add a photo (reference picture)

You can also drag and drop files onto the ① area. What happens depends on what you add:

What you add to ①What happens
🖼Photo or illustration
PNG / JPEG / WebP
It goes into a new slot (up to 9).
🔊Voice or sound
WAV / MP3 / M4A / FLAC / OGG
Added with + Add a sound in a photo's slot, it becomes that person's voice. Dropped straight onto ①, it becomes a “Sound only (no picture)” slot; you can add a photo later with + Add a picture.
🎞Video
MP4 / MOV / MKV / WebM
The “Choose a picture and a sound” screen opens. Pick one moment of the video to use as the photo and a part to use as the voice (up to 5 seconds), then press Add; both go into one slot together (untick “Add the sound too” if you do not need the sound).

②Write the video description

Write the video you want to see in the “②Video description” box. A single line is enough (for example A fantasy story). If you're not sure how to write it, start by picking an example with 💡Examples under the heading (→ Examples). To write it yourself, there are three ways.

🧭Guided writingThe easiest. Just answer questions and the description is written for you. Start here if you are new.
🪄Auto promptWrite a few words and press it. It is rewritten into English wording that suits video generation.
✍Write it yourselfPut the words you want spoken in double quotes (for example "Welcome!"). They are spoken in the language you wrote them in.
💡Examples: press the yellow 💡Examples under the heading (the first button) and pick an example. Its description and settings (and the photo or voice) are filled in. Press Create the video as is, or replace the photo or text with your own. If you have already entered photos or text, you are asked before they are replaced. The example descriptions are already formatted; Original brings back the original text before formatting. There are six examples: “New character”, “New Year 2027 (Sheep)”, “My pet talks”, “Transform: back to 16”, “Starts dancing” and “Talk with a photo and voice” (the transformation is an AI-imagined look, not the person’s real younger face). It cannot be pressed while a video is being created.
  • 🆕New character: makes a three-view sheet video (front, side and back) of the character you describe and adds its picture (the sheet) and voice to 📒Characters automatically. Choose the style and the character type from the lists (with Auto, the AI chooses). Use it from the Character book in story videos to keep the same look from any angle. → Make a three-view sheet with 🆕New character
  • 🗓Today's theme: the example screen shows a theme that changes every day (at midnight Japan time). Press Use this theme to put it in the description (add a photo to make it with that person or pet). Polish it with 🪄Auto prompt, then create.
  • 🔰Your first five: a checklist of one with a photo, one transformation, one with speech, one edited video and one photo in the album. Items are ticked automatically from the videos you have made (it does not count how many days in a row you use it).
  • 🏆Your record: shows how many videos you have made so far, how many people are in your character book, how many more videos to the next milestone, and ideas to try next.
✨Effects: press ✨Effects, to the right of 💡Examples under the heading. The effect cards are listed in color-coded sections: “AI analysis”, “Fun effects”, “Pets and plushies”, “Play a role”, “Seasonal greetings”, “For shops” and “Drawings come alive” (for example 💥Turn giant, 🗿Become a figure, 🎈Puff up, ❄️Freeze, 🦸Hero transformation, 🍰Turn into cake, 💃Starts dancing, 🤗Hug together, 🐶My pet talks, 🧸My plushie comes alive, 🥷Ninja, seasonal greetings such as 🎄Christmas, 🏪Promote your shop and 🖍️Drawing comes alive). 💞Destined partner at the top (“AI analysis”) needs no photo (see “💞AI analysis” in Handy tips). 🖐️Palm reading reads the palm photo you added in ① and turns the reading into videos told by a palm reader in a night alley (see “🖐️AI analysis” in Handy tips). 📩Otherworld Letters: your biggest ally in the world. Send a comment to someone in your 📒Character book, and their answer from another world arrives as a video letter — your heart grows lighter as you enjoy the talk (see “📩AI analysis” in Handy tips). For effects that use photos, add a picture in ① first, then press a card: a ready-made description (already in the official format) and the settings are filled in on the “Create” screen, and your pictures in ① stay as they are. Then just press Create the video. Cards marked “Type your words” ask for words first (such as your pet's line, or your shop's name and catchphrase); type them and press Use this effect. If there are not enough pictures, it tells you how many to add. Every effect makes a vertical (9:16) 6-second video (longer only when the words you typed need more time). Original shows the text the effect was made from; edit it and press 🪄Auto prompt to format it again.
  • 🤗The hug effect needs two pictures. Use it only with the consent of both people.
  • 🖍For the drawing effect, use a picture of the drawing only (do not include photos of the child who drew it). Sample → “🖍️Drawing comes alive” in 9
  • 🧸Use plush toys of commercial characters for personal fun only.

🧭How to use Guided writing

Press 🧭Guided writing under the ② heading (to the right of ✨Effects) and a screen with the following three columns opens. Answering only what you know is fine (anything can stay empty).

👤①CharactersFor each photo: age, gender, occupation, features and outfit. The age can be just a number. Tick “Non-human” for a dog, a cat and so on, and “Anime” for an anime drawing.
🎬②Scenes (scenario)Where they are and what is happening, who says what, and subtitles. Use “+ Add a scene” to add scenes in the order things happen.
📝③The descriptionThe text built from ① and ② appears right away. The box above it is for the background music of the whole video.

✨How to use Auto prompt

How to show subtitles (Guided writing)What to do
Use the spoken words as the subtitleType the words in “What are they saying?” and tick “Show subtitles” (leave the subtitle box empty).
Show a subtitle that differs from the spoken words
(a translation, narration and so on)
Type the words in “What are they saying?”, type the text you want to show in the subtitle box to the right of “Show subtitles”, and tick “Show subtitles”.
No voice, only a subtitleLeave “What are they saying?” empty, type the text in the subtitle box and tick “Show subtitles”.
No subtitlesDo not tick “Show subtitles”.

🎥Adding a camera movement

Under the description box, the “🎥Camera:” buttons (🚁From the sky, 🔄Orbit around, 🔍Move in, ↔ Slide sideways, 📹Handheld) add that camera movement to the description. Press the same button again to remove it. It also works on descriptions polished with 🪄Auto prompt.

Tip for getting what you want: results are steadier when you describe the place, the time of day, the person's movement and the spoken line one by one, in concrete terms. The quickest way is to start from the text made by “Guided writing” and adjust it little by little.

Transform the person in one photo (Guided writing)

You can make the person in the photo younger, change their gender, or turn them into a cat or an anime character. No difficult description is needed. Just fill in “①Characters” in 🧭Guided writing as shown in the table below and press ✨Format prompt.

What you wantWhat to enter in “①Characters” (example)Result
🙂Animate the photo as it isOnly Gender “male” and Occupation “Japanese”. Leave Age empty, and leave “Non-human” and “Anime” untickedThe person in the photo moves and talks just as they are
🧒Become youngerAge “16”, Gender “male”, Occupation “Japanese” (writing “high school student” in Features gets you closer)A 16-year-old look that keeps the facial features
🔄Change genderAge “16”, Gender “female”, Occupation “Japanese” (put the hairstyle and outfit in “Features” and “What are they wearing?”)A female look that keeps the facial features
🌏Change the look (appearance)Age “16”, Gender “female”, and a word that describes the look in Occupation (for example “Caucasian” or “Black”)The impression of the look changes while the facial features are kept
👶Become a babyAge “0”, plus Gender and OccupationA 0-year-old baby
👴Become 100 years oldAge “100”, plus Gender and OccupationA 100-year-old look with wrinkles and white hair
🐈Become a catTick “Non-human”, Occupation “cat”, plus Gender (and for example “calico kitten” in Features)A lifelike cat from head to toe
🐱Become an anime catTick both “Non-human” and “Anime”, Occupation “cat”A cat drawn in anime style
🎨Turn the person in the photo into animeTick “Anime” (Gender and Occupation as they are for the person in the photo)A Japanese anime character whose face resembles the person
This does not reproduce what the person really looked like. When you transform someone by entering a number in the Age box (becoming younger, changing gender and so on), the app creates a new face of the age and gender you specify, based on features in the photo such as the eyes, nose, mouth and the outline of the face. It cannot know what that person's face actually looked like at 16. Think of the result as a different person, of another age or gender, who resembles the person in the photo.
When you change the age, gender or look, always enter a number in the Age box (the person's current age is fine). If Age is empty, the video is made with the look in the photo.
To begin with, try the sample JP_Mail_50.jpg (a 50-year-old man); the change is easy to see. The detailed steps are in 9. Handy tips, “⭐With your own photo: ‘I want to look younger!’ …”.

③Add a video as a reference for motion (only when you need it)

⚠A reference video roughly doubles the video memory and time needed (it is processed together with the output, at the same size and length). When you use one, set the detail or the length to about half of what you usually use.

To use a video as the source of a “face and voice” rather than as a motion reference, add it to ① instead of ③ (→ the “What you add to ①” table above).

④Choose the detail, length and quality

At first they can stay as they are (the starting values are detail 0.9, length 1 second, shape Landscape 16:9 and quality 20, which is “Speed”). If you are not sure, just press one of the purpose buttons under “Quality” (Fast preview to Best). For details, see 4. Detail, length and quality, and how long it takes.

⑤Create the video

While a video is being created, close other apps that use the graphics card (games, video editing, video playback in a browser and so on). Leaving them open makes the video more likely to fail for lack of memory.

⏳Making several videos in a row (queue)

Press ⏳Add to queue to the right of Create the video to add the current settings to the queue. Each press adds one video, so you can make several videos one after another.

The queue is kept only while the app is open. Closing the app clears it.

4Detail, length and quality, and how long it takes

This table shows at a glance which items affect the “time (⏱)” and the “video memory (🧠)”. If a video fails for lack of memory, lower a 🧠 item; to wait less, lower a ⏱ item.

ItemMeaning⏱Time🧠Memory
DetailHow fine the picture is. Next to it you see a guide such as “HD” and the actual size in pixels. Choose from 0.2 to 2.⬆⬆higher takes much longer⬆⬆higher uses much more
LengthHow many seconds the finished video lasts. Choose from 1 to 15 seconds, or Auto: the length is chosen from the description, shown in the box on the right, and used for the video (the longer of the time needed to say the spoken lines and the time the actions and scenes need, chosen by the formatting AI; if you write time segments such as “0-1s: …”, until the end of the last one). It is chosen when you press 🪄Auto prompt or ✨Format prompt in 🧭Guided writing, or just before the video is made. With background music, use 5 seconds or more (Auto starts at 5 seconds when there is music).⬆longer takes more⬆longer uses more
ShapeThe shape of the video, such as landscape, portrait or square (Landscape 16:9, Portrait 9:16 and so on).hardly any effecthardly any effect
QualityHow carefully the picture is finished. From 1 to 50; higher is more careful (5–10 with Turbo).⬆higher takes moreabout the same
TurboSwitches to a “fast finish” for previews. Available only when the quality is 5–10. Fine details of faces and clothing become rougher.⬇⬇much fasterabout the same
SeedChooses whether the “random number the picture starts from” changes every time, stays the same as last time, or is a number you choose. There are three choices: Random, Same as last time and Custom value.no effectno effect
Reference video (③)A video used as a reference for motion.⬆about 2×⬆about 2×

On a smartphone, scroll the table above sideways to see the ⏱ and 🧠 columns.

The fast model for Turbo is installed together with the rest during setup. On a PC that still has an older setup, turning Turbo on downloads about 19.5 GB (the progress is shown next to “Turbo”).

When you add background music, set the length to 5 seconds or more

If the video is shorter than 5 seconds, it may come out silent even if you specify background music. This is because, among the official samples of the video AI (MiniMax H3), the ones with a photo, background music and spoken lines are made at 5 seconds, so 5 seconds or more is the guideline when you add background music.

Video length (seconds)

LengthWhen you add background music
🟥1–3 secondsNot recommended. The music may be left out and the video may be silent. Use this to try out the look quickly.
🟨4 secondsPossible, but there are few official samples with background music at this length, and the video may be silent.
🟩5 secondsThe most recommended. The same length as the official samples with background music.
🟩6 seconds or moreMakes it easier to combine background music, spoken lines and ambient sound.

When you use spoken lines, ambient sound and background music all together, 5–6 seconds works reliably. This caution also appears when you hover over “What about background music?” in Guided writing.

Seed (Random / Same as last time / Custom value)

Even with the same description, the framing and motion change every time you create a video. The number that decides this is called the “seed”.

To enter a number yourself: just type a number in the box to the right of Seed. It switches to “Custom value” automatically. You can also edit the number shown by “Same as last time”. While the box is empty, you cannot press Create the video.
When “Same as last time” cannot be chosen: if you have not created any video yet, or with an old history entry that has no recorded seed, you cannot choose “Same as last time” (use “Random” or “Custom value”). Once you have created one video, you can choose “Same as last time” from then on.

Even with the same seed, the result changes if you change the description, detail, length, quality, Turbo or photos. It is not for creating exactly the same video again; think of it as a way to get a similar feel.

Purpose buttons (under Quality)

ButtonWhat it setsWhen to use it
Fast previewTurbo on, quality 10When you want the fastest “preview” of the framing and motion. Not suitable for checking the finished look.
SpeedTurbo off, quality 20When you want to see a normal finish in a shorter time (this is the initial setting).
StandardTurbo off, quality 30Everyday use. A good balance between speed and how good the video looks.
HighTurbo off, quality 40When fine details, motion and a steady voice matter.
BestTurbo off, quality 50For the final finish.
Recommended flow: check the framing and motion with Fast preview → adjust the description → when you like it, re-create it with Standard or higher. 🎬Make it in full quality in History fills in the same description, materials and seed with the quality set to Standard (→ Make it in full quality).

How long it takes

At the very bottom of the settings (under the Seed box) you see how long it will take. There are two kinds of estimate.

🆕First-run estimateShown while this PC has no record yet. A rough figure calculated from the length and quality only (see the diagram below).
📈On this PCAfter one video it switches to this (with some graphics cards it appears from the first video). It is based on the time videos actually took, so it is more accurate. Example: “On this PC: about 6 min 50 sec (from 3 past runs with the same settings)”
Example of the first-run estimate: a length of 3 seconds gives three blocks of 2 min 10 s, quality 20 gives two, five in total, about 10 min 45 s ⏱Length 3 s (one per second) ✨Quality 20 (one per 10) Total 2 min 10 s 2 min 10 s 2 min 10 s 2 min 10 s 2 min 10 s = about 10 min 45 s
The first-run estimate is a sum: about 2 min 10 s per second of length + about 2 min 10 s per 10 of quality

5Using History

The “History” tab lists the videos you have created, newest first. Type a photo's file name, part of a description or a creation date (for example 2026/9/9) in the search box at the top to narrow the list. When you have many videos, more appear as you scroll down (or press Show more at the bottom). A search looks through all of your history.

Video titles: each video shows a short title at the top (up to 40 characters in English, or 20 in Japanese). The formatting AI makes it from the prompt automatically (when you rewrite the prompt with 🪄Auto prompt and similar buttons, or, if there is no title yet, right after you press Create the video or ⏳Add to queue). It can take a little while; to skip waiting, press Continue without a title. Once a video has a title, small changes to the prompt do not make a new one. Videos without a title (such as those made with an older version) show a summary of the prompt in light text. The same title also appears in the queue list, Thumbnails, Edit video and the album. To change it, use ☰Menu → ✏️Edit title.

ButtonWhat it does
Click the thumbnail (the picture with a ▶ mark)Plays the video.
Works marked “No video”The record is still there, but the video file is missing (it may have been moved or deleted). You can make it again with 🔁Make it again in ☰Menu.
🖼ThumbnailsPick your favorite moment from the video. You can use that moment as the cover (thumbnail), or keep it as a photo with ⭐Save to album.
🎬Edit videoFound inside Thumbnails. Hide faces and other parts, lay mask images over it or delete frames you do not need, and save a video for sharing as a separate file (the original video stays as it is). For details, see “Edit video” below. You cannot press it while a video is being created (editing also uses the graphics card, so running both at once makes the generation more likely to fail).
📁Save folderOpens the folder that holds the video.
📤ShareSends the video to apps such as LINE or Instagram. Press 📤Choose an app and send and the share screen opens, where you pick the app. The post text includes “#AIgenerated” and the link to the WishMallet page, and you can edit it before sending (when you post an AI-made video, make it clear that it was made with AI). If it cannot send directly to apps, 📋Copy the text and 📁Save folder appear instead: open the video's folder and send it from the app. It is also available on 🛡For sharing rows (videos made with “Edit video”).
View inputsShows a list of the photos, voices and reference videos used at that time. When only part of a long reference video was cut out and used, it also shows the length actually used and where in the original video it starts.
🔁Make it againBrings the same photos, description and settings back to the “Create” screen. What you typed in Guided writing comes back too. If the description was rewritten with 🪄Auto prompt, Original brings back the text from before (what you typed yourself). Handy when you want to change a little and create again. The seed carries over the value that was actually used at that time, and “Same as last time” is selected from the start (switch back to “Random” if you want to change it).
🔁Load these settingsBrings settings you kept with “💾Save to history” back to the “Create” screen. What you typed in Guided writing comes back too.
🔧AdjustPick one change (🏃Bigger movements, 😊Smiling, 🔍Closer camera, 🐢Slower, ☀Brighter or 🌙Make it night). It is added to the description and the “Create” screen is filled in with the same seed, so you only need to press Create the video. For a formatted description, the same instruction is also added to the original text that Original shows.
🎬Make it in full qualityAppears only on videos made with “Fast preview”. Fills in the same description, materials and seed on the “Create” screen with the quality set to Standard (no Turbo). Fast preview uses a different model, so the result looks similar rather than identical.
🎨Remix the lookKeeps the motion and composition and makes the video again with a different look. Press a look (🌙Night background, ⛄Winter scenery, 🎨Watercolor, 🧱Clay animation, 📼Old film or 🖌Oil painting): the video goes into ③ as the reference video, an instruction is added to the description, and the “Create” screen is filled in (for a formatted description, the same instruction is also added to the original text that Original shows). Because it uses a reference video, it takes about twice the memory and time.
▶Make the next partPuts the last scene of the video into ① as the photo and fills in the “Create” screen with the same settings (voices in ① are not carried over, because the person in the last scene may not have the same voice; add a voice again if needed). Rewrite the description for what happens next, then press Create the video.
📤ExportWrites out a file for trying the same settings on another PC or in the cloud (for advanced users → 7. Export).
Copy descriptionCopies the description used at that time.
Merge split videosJoins the videos made from a split reference video into one video, in the order they were split (shown only for videos made that way). Pressing it opens a screen where you check the videos and their order. Click a thumbnail to play that video. Uncheck the videos you want to leave out, and drag a row to change the order (the count and total length cover only the checked videos). When a video has versions made with “Edit video” (shown as “For sharing” in History), ◀ ▶ appear beside its thumbnail so you can choose which file to join, in the order “Original → Edited 1 → Edited 2 → Original” (this lets you join a video with frames deleted or places hidden; clicking the thumbnail plays the chosen file). When you join a video that became shorter because frames were deleted, “Join the sound of the generated videos” keeps the picture and sound together better. For the sound, choose “Use the sound of the original reference video” (there are no joins in the sound, so it suits music and dance videos) or “Join the sound of the generated videos” (each join is blended smoothly). Set the file name, press Save, and choose where to save it; merging then starts and shows its progress in %. If you deleted some of the videos, only the remaining ones are joined. If some parts were deleted or not made yet, press Queue the ○ missing parts (○ is the number of missing parts) to queue just those parts with the same settings. When they are done, open Merge split videos again. The finished video is saved where you chose and is also added to History. To remake just one of the parts, use 🔁Make it again on that video; the newer one is used when merging.
✏️Edit titleChanges the video's title yourself (up to 40 characters). Save it empty to show a summary of the prompt instead.
Delete entryDeletes that history entry. On the confirmation screen, choose “Remove from list” (removes it from the list only; the video file stays, so you can bring it back later from “Settings” → “History”) or “Delete the video too” (deletes the whole save folder; this cannot be undone).
Delete in a 🛡For sharing rowVideos for sharing made with “Edit video” are listed one row each under the original video. You can delete just that one item (the original video is not deleted).

A failed entry keeps its “Failure details”. They are a useful clue when you contact support.

Edit video (hide faces, lay mask images, delete frames you do not need)

This screen hides “parts you do not want to show” in a finished video and removes scenes you do not need, to create one more video for sharing. The original video is not changed. All editing happens inside that PC only, and no video or image is sent anywhere.

How to open it: in “History”, press 🖼Thumbnails for the video, then press 🎬Edit video at the bottom left.

From top to bottom the screen has three levels: ①the video (the dotted oval is a “hidden place”) → ②the mask image strip and the frame strip (the upper row holds mask images laid in front of the video; the lower row is the video's frames laid out side by side, and striped frames are “frames to delete”) → ③the row of buttons. Pick a moment on the strip, check it in the video, and edit with the buttons. At the very bottom are 🛡Privacy blur, 🔞Review mosaic and Save video, with the note “AI can miss things” below them.

The first time you use “Sensitive body parts” or 🔞Review mosaic, a screen asks to download detection data (about 12 MB). Press I have read the license, fetch it and it is ready from then on (the detection data for faces, people and license plates is included from the start). Your videos and images are never sent anywhere.

Step 1: Watch the video and pick a moment

What you wantHow
Play / pauseClick on the video to switch between play and pause each time (▶and ❚❚marks appear in the middle of the screen). The Space key does the same.
Go to the moment you want to seeClick the frame strip (or drag it). The moment you clicked appears in the video.
Move just a little← → move one frame at a time; hold Shift as well to move 0.5 seconds at a time.
See where things are hiddenTick ☑Show hidden places to show outlines around the hidden places (it is ticked from the start). Untick it to see exactly how the published video will look.
Compare with the unedited videoSwitch between Processed and Original to check the original video too.

Step 2: Hide what you do not want to show

There are three ways to hide things. You can combine all of them (adding one later does not remove the earlier edits).

MethodHow to do it
🛡Privacy blur
(finds places automatically)
In “What should be hidden?” choose Faces / Whole people / Sensitive body parts / License plates, set the Strength (Standard, Strong, Maximum) and press Auto process. It looks through the whole video and finds the places to hide automatically. To change how they are hidden (Blur, Mosaic, Black box), choose it under “Show advanced settings”.
🔞Review mosaic
(for stores)
Choose the store (DLsite / FANZA Doujin) and the Strength (Recommended, Stronger) and press Auto process. The groin, anus and contact areas are hidden automatically. To hide the chest and buttocks too, add them under “Additional parts”. This is automatic processing that refers to each service's publishing standards; it does not guarantee that the video passes review.
Hide this too
(mark it yourself)
Press the button, then drag on the video to draw an oval around the place. Choose Hide it for (1 second around this moment / The whole video) and press Hide this place. Use it for places that were not found automatically.

You can review the hidden places with the following two drop-down lists.

Drop-down listWhat it does
Hidden places: NLists the places being hidden now (Face 1, License plate 2 and so on). Pick one to jump to the moment where that place first appears. For an edit you do not need, press Delete on the right to stop hiding it throughout the video (you can bring it back with “Undo”, which appears on the screen). You can do the same by clicking the outline in the video directly and pressing Remove this mask.
Scenes to check: NScenes where automatic detection judged that “something may not be fully hidden” (tracking stopped partway, it touches the edge of the frame, it appeared only for an instant and so on). Pick one to jump to that scene, then look and check with your own eyes. If there is no problem, press Delete to remove it from the list (the hiding stays as it is; this only marks the scene as checked).

Step 3: Delete frames you do not need

Things that got into the shot by accident and moments that went wrong can be removed frame by frame.

What you wantHow
Delete the frame you are looking atPress 🗑Delete this frame. The Delete key also works (while a mask image is selected, Delete removes that mask image instead; click the frame strip to go back to deleting frames).
Delete a run of frames at once (range)Pick the first frame of the range, then right-click the frame strip (long-press on a phone) and choose Select a range from the menu. Now drag on the strip: everything from the first frame to where you drag is selected in blue (you can also drag to the left). Press 🗑Delete range to delete all the selected frames at once (the Delete key also works). To cancel, press Esc or choose Cancel the range from the menu.
Bring back a deleted frameGo to that frame and press ↩ Keep this frame (pressing Delete again also brings it back). To bring them all back at once, right-click the frame strip (long-press on a phone) and choose Undo all deletions from the menu.
See how much you deletedDeleted frames are marked on the strip, and the number and length are shown, like “Frames to delete: 12 (0.5 s)”.

Step 4: Lay mask images over the video (cinemagraph / motion comic)

What motion comics are good for and tips for making them are in 🎞Make a motion comic.

When you lay a partly transparent image (a mask image) in front of the video, the video moves only in the transparent parts. You can make a comic page where “only the panel being read moves” (a motion comic), or a photo where only one part moves (a cinemagraph).

How mask images work: lay a comic page with one panel made transparent in front of the video, and the video behind shows and moves only in that panel ▶ ▶ + Mask image (front)a panel made clear Video (back)moves as usual Resultonly that panel moves
The mask image sits in front of the video. The video behind shows only through the transparent parts
What you wantHow
Add a mask imageClick the dotted + Add a mask image (an empty space) just above the frame strip and pick an image. It fills that whole empty range (one image alone covers the whole video).
Line up one image per panelPick several images at once in the file window, and they are placed in file-name order (01, 02, 03…), splitting the empty range evenly. For a comic, prepare images with one panel made transparent each, in reading order, and those panels move one after another.
Make a mask image from this framePick a moment and press 🖼Make a mask image from this frame to open the “Make a mask image” screen. Drag to surround the area you want to move (⬭Oval / ▭Rectangle), then press Make this area transparent (as many areas as you like; ↩ Undo one brings the last one back). Save turns it into a PNG and puts it in the last empty space of the strip (if there is none, the last mask image is halved and the new one goes after it; if the strip is empty, it starts at the beginning). Pressing ✕Close before saving makes nothing.
Shift when it is shownDrag the mask image. When it bumps into another mask image, it pushes that one along instead of overlapping. Once they are packed against the start or end of the video, it stops. After selecting one, ← → also move it one frame at a time (0.5 seconds with Shift).
Change how long it is shownDrag the left or right edge of the mask image (mask images it bumps into are pushed along).
Change the imageDouble-click the mask image, or select it and press Change image (the range stays the same).
Remove itClick the mask image to select it, then press the Delete key or Delete. Only the selected mask image is removed (use Undo in the notice to bring it back).
Remake or remove it (right-click or press and hold)Right-click (on a touch screen, press and hold) a mask image on the strip to get ✏Edit and 🗑Delete. “Edit” opens that mask image on the “Make a mask image” screen so you can add more transparent areas (saving replaces it with a new PNG, and the range stays the same). “Delete” works like the Delete button.
Check how it looksIn the “Processed” view, the mask images are laid over the video both when it is paused and while it plays (the “Original” view does not show them). On the strip, the panels made transparent are painted purple.

Step 5: Save the edited video

AI can miss things. Before you publish a video, always watch it yourself from start to finish and check it.

🎞Make a motion comic (bring comic panels to life one at a time)

A “motion comic” is a video of a comic page where only the character in the panel being read moves and talks. With this app, you can turn a single comic page into a vertical short video where the panels come to life one after another.

📱A popular short-video formatIt is one of the “formats that can go viral” for vertical shorts on TikTok, YouTube Shorts and Instagram Reels.
📈A million views is not rareComic-style short videos passing 1 million views are not unusual. Companies also use them in ads and promotions.
🔥Anime and manga are huge on TikTokFor some works, related videos are viewed hundreds of millions to over a billion times.

Why use mask images?

Even if you ask the video AI (MiniMax H3) to “move only this panel”, when the same character's face appears several times on the page, the faces in other panels may move or lip-sync too. To the AI, every face looks like the same person.

That is where mask images help. When you lay a “still picture” with only the panel you want to move made transparent over the video, the video shows only inside that transparent panel and the other panels stay still. Prepare one mask image per panel, line them up in order, and you have a motion comic whose panels move one at a time.

How mask images work: without masks, the same character's face moves in two panels at once; with a mask image laid over the video, only the panel made transparent moves lay a mask image ❌Without masks other faces move too ✅With masks only one panel moves
The striped areas are the mask image (a still picture). Only the panel with the green frame shows the video behind and moves

How to make one (4 steps)

①Make a video from a comic pagePut the comic page in “①Reference pictures” and choose “Portrait 9:16 (phone)”. Describe the movements, expressions and lines in the prompt
②Open Edit video“History” → 🖼Thumbnails → 🎬Edit video
③Make a mask image per panel“🖼Make a mask image from this frame” → surround the panel to move → “Make this area transparent” → Save. Repeat for each panel in reading order
④Check, save and postEven out the mask lengths on the strip and click the video to play it. After “Save video”, post it as a short!
TipsDetails
📚Follow the reading orderJapanese comics are read from the top row to the bottom row, and right to left within a row. Make (and line up) the mask images in this order so the story flows naturally.
📷Keep the camera stillIf the whole page zooms or shakes, the still picture (the mask image) and the video no longer line up. Writing “camera fixed” in the prompt helps.
⏱About 2 to 3 seconds per panelFor a 10-second video with 4 panels, that is about 2.5 seconds each. New mask images go to the end of the strip (if there is no room, the last one is halved to make space), so finally drag the edges on the strip to even out the lengths.
🗂Add image files at onceIf you have PNGs with one panel made transparent (one per panel), pick several at once with the dotted + Add a mask image on the strip, and they are lined up in file-name order, split evenly.
✏Remake with a right-clickRight-click (on a touch screen, press and hold) a mask image on the strip → ✏Edit to add more transparent areas. 🗑Delete removes only that mask image.
💾Safe to close midwayMask images are remembered for each video. Close the screen, and next time you open “Edit video” you can continue where you left off.

For step-by-step operations, see “Step 4: Lay mask images over the video” in 5. Using History.

Use only comics and illustrations that you own the rights to or are allowed to use (→ 11. Using the app safely).

6Best shots

A place to collect “favorite single frames” from your videos as photos. Photos are added with 🖼Thumbnails → ⭐Save to album in “History”.

💾Save full quality (enlarge with AI for your purpose)

Press 💾Save full quality on a photo in the large view, and a screen asks “What will you use it for?” so you can pick a purpose. Just pick one: the AI enlarges that moment of the original video to just the right size and saves it as a lossless PNG.

①Pick a purpose“Kindle e-book cover”, “Poster” and more. Not sure? Pick “⭐Master image”
②Choose the areaDrag the bright frame in the preview to choose the part to keep
③SavePress “💾Save” and choose where to save; a 0–100% progress bar appears
④DoneCheck the image with “📁Open folder”
How Save full quality works: crop one frame of the original video to the shape of the purpose, enlarge it 4× with AI, and save a PNG at the exact size 🎬🤖🖼 Crop the framefrom the video AI enlarges it 4×cleanly Save a PNG atthe exact size
The frame is taken from the original video without quality loss and enlarged with AI. Images for books and posters are saved at 300 PPI for printing
PurposeSize (pixels)Use it for
⭐Master image (use it for anything)7200 × 10800Not sure? Pick this. Saved at the largest size, so you can cut covers and posters from it later
📱Kindle e-book cover1600 × 2560The cover shown in the Kindle store
📕Paperback cover (5 × 8 in)1538 × 2475Novels and smaller books
📗Paperback cover (5.5 × 8.5 in)1688 × 2625Novels and general books
📘Paperback cover (6 × 9 in)1838 × 2775The most common size on Amazon KDP (printed books)
📙Paperback cover (7 × 10 in)2138 × 3075Practical, photo and illustration books
🖼Poster (2:3, up to 24 × 36 in)7200 × 10800Selling posters, photos and art
🖼Framed art (3:4, up to 18 × 24 in)5400 × 7200Art and posters to frame
🖼Wall art (4:5, up to 16 × 20 in)4800 × 6000Illustrations and interior decor
🖼A-series poster (A4 to A1)7016 × 9933Print and sell in A4, A3, A2 and A1
📷Original size (no enlargement)Same as the videoThe video frame as it is, as a lossless PNG

The sizes are for portrait. For the master image and the posters and art, you can also choose landscape (for example 10800 × 7200) under “Orientation” (for a landscape video, it starts as landscape).

🎁During the free trial you can still open 💾Save full quality and see the list of purposes and the preview. When you press the final 💾Save, it shows how to get the full version. 💾Save all as ZIP is also a full-version feature. Saving to the album and viewing it work.

A quick way to make one high-quality photo is in 9. Handy tips.

7Export (for advanced users)

The instructions and messages in exported files follow the Japanese or English language selected in Settings. Your prompts, material names and file names are kept as entered.

This chapter is for people who want to re-create a video on another PC or on a cloud service. You do not need it for everyday use.

📤Export is a button that writes the settings you are using now (description, which photo is which, detail, length, quality and so on) into one file. Open the file in another environment and you can re-create the video with the same settings. It is not a button that exports the video itself (the video is in your save folder).

🎁During the free trial you can open the export screen and check its contents. Export (writing the file) is a full-version feature.

The photo, voice and video files themselves are not included. Only the file names are recorded. In the other environment, add the files with the same names, as in the list shown on the screen. If a name is different, the file cannot be loaded.

Three kinds of files

FormatWhen to use it
🖥①For ComfyUI (standard)To open it in ComfyUI installed on your own PC. It works the same way as this app, so the result is the closest.
☁️②For ComfyUI (high quality)To re-create the video at a higher quality on the cloud service Floyo.
🌐③For Google ColabWhen your own PC has no high-performance graphics card. It runs on Google's cloud and does everything automatically, from setup to creating the video.

When you open the Export window, “③For Google Colab” is selected at first (to export for ComfyUI, choose another one under “Format”).

How to use them

  1. ①ComfyUI (standard): copy the exported .json to the PC with ComfyUI, and put the photos and voices into ComfyUI's input folder with the same names. Open ComfyUI in a browser, drag and drop the .json onto the screen and run it.
  2. ②ComfyUI (high quality): open the ComfyUI screen on Floyo, drag and drop the .json, upload the photos and voices, and run it. If “Missing Models” appears, follow Floyo's guidance to get the parts.
  3. ③Google Colab: in Google Colab, load the .ipynb from “File → Upload notebook”, choose a GPU (G4 is the best choice; if you cannot select it, A100) under “Runtime → Change runtime type”, turn on “High-RAM” if the option is there, and choose “Run all”. When it pauses partway, upload the photos and voices with the names shown. When it finishes, the video is downloaded to your PC automatically.

Choosing a GPU in Google Colab

A guide to which of the GPUs offered under “Change runtime type” suit this app. For video creation, the amount of video memory matters most (→ How do they compare with home cards?).

GPUFitWhy
G4◎Best choiceAbout 96 GB of video memory (NVIDIA RTX PRO 6000 Blackwell). The full set of AI parts (about 42.5 GB) fits as it is, so the waits are the shortest.
Cost guide: Colab’s G4 (96 GB) is about $0.87 per hour. You are charged only for the time you use.
A100◎Powerful and stable. However, on Colab it can be the 40 GB model, which has less video memory than G4, so some swapping happens.
Cost guide: Colab’s A100 is about $0.676 per hour. You are charged only for the time you use.
L4○24 GB. It works, but there is a lot of swapping, so it takes time.
T4△Not recommended16 GB. The standard T4 stops for lack of memory. Even with High-RAM it is very tight and takes a very long time.
You can start for free. Google Colab needs no sign-up if you have a Google account (however, G4, A100 and L4 are GPUs you can select on a paid plan such as Colab Pro. The free tier usually offers only the T4, which is not suited to creating videos). For Floyo, enter KATORI30 in the Promo Code box on the payment screen to try it free for 30 days. → 9. Handy tips

How do cloud graphics cards compare with home ones?

This helps you decide when “my own PC is too slow” or “I am not sure whether to buy a new graphics card”. The first thing to know is that what matters in video creation is “the amount of video memory” rather than speed.

The full set of AI parts this app uses is about 42.5 GB even in its smallest configuration. Whatever does not fit in video memory is calculated while being swapped to and from the PC's main memory each time, and that waiting makes it slower. Even the top home card has 32 GB of video memory, so a home graphics card cannot hold the full set of parts. This is the biggest difference from the cloud.

EnvironmentVideo memoryCompared with 42.5 GBWhat happens
Your own PC (this app's minimum)12 GBabout 30.5 GB shortIt runs, but with a lot of swapping, so it takes time. With a length of 1 second and one photo, you can try up to Detail 2 (1888x1088px: Full HD).
GeForce RTX 4090 (a high-end home card)24 GBabout 18.5 GB shortMuch faster, but the swapping remains.
GeForce RTX 5090 (the top home card)32 GBabout 10.5 GB shortAmong the strongest for home use. Even so, not everything fits.
Google Colab (L4)24 GBabout 18.5 GB shortThe same amount of video memory as the RTX 4090. However, its memory reads and writes at about one third of the 4090's speed, so the waits are longer.
Google Colab (A100 40 GB)40 GBabout 2.5 GB shortComes close to fitting everything.
Google Colab (A100 80 GB)80 GBabout 37.5 GB to spareThe full set of parts fits as it is. Hardly any swapping happens.
Google Colab (G4)96 GBabout 53.5 GB to spareThe most headroom among the GPUs Colab offers. It is an NVIDIA RTX PRO 6000 Blackwell, with the video memory of 3 RTX 5090 cards.
Floyo (H100 NVL)94 GBabout 51.5 GB to spareAs much headroom as G4. Its video memory equals about 2.9 RTX 5090 cards, and its memory reads and writes about 2.2 times as fast.

On a smartphone, you can scroll the table above sideways.

The larger the detail, the longer the video and the more photos you use, the more this difference shows. In order, they rank as follows.

RTX 4090 < RTX 5090 < Colab (A100 40 GB) < Colab (A100 80 GB) < Colab (G4) / Floyo (H100 NVL)

This does not mean “Floyo is 3 times as fast as the RTX 5090”. The figure of 2.9 cards compares the amount of video memory. The actual speed also varies with the detail, length and quality, and with how busy the service is that day. It is also a different thing from a ranking of gaming performance.

Which one should you choose?

💡Could a GPU actually cost you less?
If the cloud charges JPY 100 per 10 seconds, 10 clips a day adds up to JPY 30,000 a month. On your own GPU, each clip costs only a few yen in electricity. For a few clips a month the cloud is fine, but the more you create, the more a GPU becomes the cheapest choice. → Could an expensive GPU cost less? (calculate with your own numbers)

Sources of the figures: the 42.5 GB full set of parts is from the official ComfyUI blog; Floyo's H100 NVL (94 GB, 3.9 TB/s) is from Floyo's official information; the RTX 5090 (32 GB, 1.792 TB/s), A100 and L4 (24 GB, 300 GB/s) are from NVIDIA's official specifications, and Colab's G4 (NVIDIA RTX PRO 6000 Blackwell, 96 GB) is from Google Cloud's official specifications (as of October 2026). Besides video memory, there are also differences in how AI calculations are done and in hardware generations, so please treat the comparison above as a guide.

When Google Colab stops or shows an error

Even if an error appears in red, the work done up to that point is kept. You do not need to start over. Press the ▶ to the left of where it stopped and run just that block (cell) again.

What you see on the screenWhat to do
NVIDIA GPU is unavailable.In the menu at the top, choose “Runtime” → “Change runtime type”, select a GPU (G4 is recommended; otherwise A100), and choose “Run all” again (→ Choosing a GPU).
Insufficient system RAMOn the same screen, turn on High-RAM or change the GPU to G4 or A100. The standard T4 is not enough.
It stops at the upload of photos or voicesChoose a file with the same name as the file name shown on the screen. You cannot go on if the name is different. Run the cell again to choose again.
Downloading the parts (AI models) stopped partwayRun the 04_Download_MiniMax_H3_models cell again. It skips what has already been downloaded correctly and fetches only what is missing.
“could not fetch”, “HEAD … does not match” or “installed versions differ”: a version mismatchUpdate this app to the latest version, then export again. An exported file remembers the versions of the parts that were verified to work at the time it was made, and is built not to switch to newer versions on its own (that way the same result can be re-created).
After you left it for a while, it went back to the initial stateGoogle Colab disconnects after a certain time and its contents are erased. Choose “Run all” again (it starts over from downloading the parts).

At the very bottom of the notebook there is also 10_Troubleshooting. Please take a look there first too.

If it still does not work (contacting support)

Prepare the following three things and send them to the support contact written in your purchase email. The cause can then be found quickly.

  1. The error text: copy the last few lines of the red text (the line that starts with “❌” and the lines below it are enough). A photo of the screen is fine too.
  2. The run log: press the folder icon at the left edge of Google Colab, right-click wishmallet_studio_export_report.txt and download it. The log up to the point where it stopped is saved automatically.
  3. The exported file: the .ipynb you used, itself. It records when and with which version it was made.

If the app itself also seems to be misbehaving, please also send the ZIP made with Save support information under “Settings” → “Support” (→ 10. Troubleshooting). It does not include photos, voices, videos, descriptions or your activation code.

Exported files go into the exports folder inside your save folder. In another environment the graphics card is different, so the picture will not be exactly the same. Lowering the length or detail makes failures less likely.

8The Settings screen

The “⚙️Settings” screen (the button at the right end of the tabs) lists the following, from top to bottom.

🔄Updates
When “Check for updates at startup” is ticked, the app checks for a new version every time it starts (it is ticked at first). Check now checks by hand.
🌐Language / 表示言語
Switches between Japanese and English. It changes the moment you choose. This page, which opens from ❓Help and 💡Handy tips, and the check tool that opens from Check PC performance are also shown in the language you chose.
🔔Sound effects
When “Play sound effects” is ticked, a short sound plays once when a video is finished or “Format prompt” ends (a two-note “ding-dong” on success, and on failure), and when something needs your check. It is ticked at first. Untick it and no sound plays. 🔔Test sound lets you check right away that you can hear it.
💬Pop-up messages
When something does not work (for example, you chose a file type that can't be used, or the queue is full), a light red message appears at the bottom of the screen. Choose how long it stays in “Hide automatically after” (2.5 seconds at first). 💬Show a sample shows an example. Press × to hide it at once.
✨LLM used for formatting
Chooses the AI that Auto prompt and Format prompt use. The default is the built-in LLM that runs on this PC; you can also switch to OpenRouter (→ Format with OpenRouter (sign-up and creating an API key)).
🔑License
Shows the registration status on this PC and the registered email address. One activation code can be used on one PC. If it shows “Not registered”, or if you want to change it, contact the support address written in your purchase email. During the free trial it shows “Free trial” (days left and videos today) with Enter a code and Buy the full version (→ Free trial).
💻Your PC
Information about the graphics card and memory. Check PC performance gives a score that shows how comfortably this PC runs the app.
💽Storage
Free space, and the space used by videos and temporary files. Clean up temporary files deletes files that are no longer needed (finished videos and the history are kept).
🔍Detection data (video editing)
The data for finding sensitive parts, fetched for 🔞Review mosaic. It shows whether it is fetched and its size. Delete removes it (it is fetched again the next time you use it).
📁Save location
The folder where finished videos go (at first Videos\MiniMaxH3). Change picks another folder and Open save folder opens it. Inside it, samples holds the sample photos and voices, docs the guide, input copies of the files you added, exports exported files and bestshots the album photos (samples and docs come back when you restart the app if you delete them).
⏸Auto shutdown of the video engine
If a certain time passes without creating a video, the mechanism running in the background is stopped automatically to free the graphics card's memory. When you want to use the graphics card for another app, press Stop now.
🩺Diagnostics
Check the video engine checks whether the PC is ready to create videos. When it says “Action needed”, what is missing and how to fix it are shown as numbered steps. If something is missing, press 🛠Run setup to start preparing everything for video creation (WSL2 and the video engine) in a black window; it sets up only what is missing (if Windows needs to restart, the black window tells you, and the setup continues by itself afterwards). It cannot be used while a video is being created, a description is being formatted or a picture is being saved in full quality.
🗂History
Restore history from the save folder restores videos that disappeared from the list, using the save folder.
📮Support
When you are in trouble, press Save support information to make a file (ZIP) for investigation. It does not include photos, voices, videos, descriptions or your activation code.
ℹ️About
The current version, Open the guide (opens Read me first.html, the getting-started guide in your browser) and Take the tour (shows the short first-start walkthrough again).
🤝Affiliate program
Apply for the program that pays you a commission for introducing this app, and see your current status (not applied / pending / registered rate). See “Affiliate program” below for details.

🎁Free trial

If you start without an activation code, you get a 7-day free trial. The days left and today's count are always shown at the top right.

What you can doWhat you cannot do (full version only)
✅Create videos: up to 3 a day (the count resets every day)
✅Guided writing and Auto prompt
✅History, saving to the album, checking edits in Edit video
🔒Save video (Edit video)
🔒💾Save full quality and 💾Save all as ZIP (Best shots)
🔒Export
🔒Use this in ✨Effects AI analysis (the analysis and report are included)

Affiliate program (introduce this app and earn a commission)

Introduce this app on social media, a blog or in videos, and earn a commission on every sale made through your own link. Sales and payouts are handled by Gumroad (the store). You apply from “Affiliate program” at the bottom of Settings.

RankSales so farCommission (on a 9,600 yen sale)
START0–910% (960 yen per sale)
PARTNER10–9930% (2,880 yen per sale)
TOP PARTNER100 or more50% (4,800 yen per sale)

Amounts are “price × rate” as a guide. What you actually receive changes with Gumroad's fees, refunds and so on. At the 10% rank, 10 sales add up to the price of this app (9,600 yen).

How to apply

  1. Create a Gumroad account and finish its payout settings.
  2. In the app, open Settings → “Affiliate program” and press Apply (available once you have registered a full-version license; during the free trial you can only read about the program).
  3. Enter your Gumroad account email, a contact email and your name or handle, then press Submit. Your status becomes “Application pending”.
  4. When we have registered you on Gumroad, Gumroad sends you your affiliate link and the app shows “Registered (10% commission)”.
  5. When you reach 10 or 100 sales, press Request a rank-up in the same place. We check the Gumroad sales records and change your rate to 30% or 50%.

Your current status (not applied / pending / registered rate) is checked automatically each time the app starts. Press Refresh status to check right away. If you typed the wrong Gumroad email, you can correct it with Edit application while the status is “Application pending”.

Program rules
  • Refunded or cancelled orders, your own purchases and fraudulent orders do not count as sales.
  • Commissions are paid under Gumroad's system and terms.
  • Rate changes are decided by the seller after checking the Gumroad sales records.
  • Untrue or misleading promotion, or abusive behaviour, can lead to suspension.
  • When you use videos made with this app for promotion, follow the notes in “Using the app safely” (such as getting permission from the people shown).

Questions about the affiliate program: affiliate@aicouturelab.com

🌐Format with OpenRouter (sign-up and creating an API key)

By default, 🪄Auto prompt and “✨Format prompt” in 🧭Guided writing use an AI that runs inside your PC (the built-in LLM) to polish your description. If you switch to OpenRouter in Settings, formatting becomes much faster. This chapter walks you through everything, from signing up to setting up the app, so that even a first-time user can follow it.

You can keep creating videos exactly as before without OpenRouter. You only need this chapter if you want to format with OpenRouter.

①What is OpenRouter? What you need, and how it differs from the built-in LLM

OpenRouter is an internet service that lets you use many different AIs through a single door (it is run by a company other than the maker of this app). This app asks one of them, deepseek/deepseek-v4.1-flash-1017 (DeepSeek V4.1 Flash 1017), to polish your description.

What you need

How it differs from the built-in LLM (the default)

ItemBuilt-in LLM (default)OpenRouter
Speed of formattingAbout 30 seconds to 2 minutesFast
Where your description goesStays inside this PC. It never leaves itSent over the internet to OpenRouter and to the provider that runs the model (→ ⑦)
CostFreeA small fee applies (about $0.0007 per run, as a guide)
InternetNot needed for formattingNeeded every time you format
PreparationNone (only the one-time download of parts)Sign-up, credits and an API key (→ ②–⑤)
Cost guide: about $0.0007 per run (a small fraction of a US cent). This is a guide. It goes up for long descriptions or many scenes, and it also changes with the provider that runs the model and with the time. For the latest prices, see the model page on OpenRouter (https://openrouter.ai/deepseek/deepseek-v4.1-flash-1017). What you have actually spent is shown in OpenRouter's Activity (https://openrouter.ai/activity).

②Sign up for an OpenRouter account

  1. In your browser, open https://openrouter.ai/.
  2. Press Sign in at the top right of the page.
  3. Choose whichever is easiest for you: Google, GitHub or an email address (if you have a Google account, that is the easiest).
  4. Follow the on-screen instructions. If you sign up with an email address, also confirm the email that is sent to you.
  5. When the OpenRouter pages open while you are logged in, you are done.

③Add credits

With OpenRouter you buy credits (a balance) first, and only what you use is deducted. Without credits, formatting stops with “not enough credits”.

  1. Open https://openrouter.ai/settings/credits (if you are not logged in, the login page appears first).
  2. If you pay by credit card, we recommend adding $15 or more in one purchase. A fee is charged on every credit purchase, so one purchase of $15 or more costs less than several small ones (see the note below). How many runs $15 covers (a guide): formatting costs about $0.0007 per run, so $15 covers about 20,000 runs on average. With short descriptions only, that is about 28,000 runs; even with long descriptions of about 3,000 characters every time, it is about 9,000 runs. Long descriptions or many scenes mean fewer runs, and the number also changes with the provider and with the time. When it runs low, just add more.
  3. Enter a payment method (such as a credit card; choose from the ones OpenRouter shows) and complete the payment.
  4. When the balance has gone up, you are done.

About the fee: a fee is charged on every credit purchase. For credit cards it is 5.5% of the amount (with a $0.80 minimum), according to OpenRouter's FAQ (as of October 2026). Because of the minimum, a purchase under about $14.55 pays a higher percentage than 5.5%. From $15 up the rate stays at 5.5%, so with a card, buying $15 or more at a time is the better deal. Fees and payment methods can change, so what OpenRouter's own pages show is the correct information.

④Create an API key

  1. Open https://openrouter.ai/settings/keys.
  2. Press Create API Key.
  3. In Name, enter a name you will recognize later (for example: WishMallet).
  4. We recommend setting a limit (Credit limit). With a limit, even if someone else used your key by mistake, no more than that amount could be spent.
  5. Press the button that creates the key.
  6. A long string that starts with sk-or-v1- (the API key) appears. Copy it, for example with the copy button.
This is the only screen where you can see the whole API key. Once you close it, you can never see the whole key again (if that happens, create a new key). As soon as you have copied it, paste it into the app in step ⑤ below.
An API key is like a password. Do not show it to anyone, do not post it on social media or a blog, do not let it appear in screenshots or videos, and do not send it to other people by email or chat. If you think it may have leaked, delete that key on the keys page above and create a new one.

⑤Set it up in the app

  1. Press ⚙️Settings (the button to the right of ❓Help) and find “✨LLM used for formatting”.
  2. In “Type”, choose OpenRouter (deepseek/deepseek-v4.1-flash-1017).
  3. Paste the key you copied in ④ into the API key box (Ctrl + V).
  4. Press Save. The saved key is shown after “Saved:”, and the box is emptied.
  5. Press Test connection to check that it connects. When it says “Connected”, you are ready (it does not format anything, so it costs nothing).
  6. From now on, press 🪄Auto prompt or ✨Format prompt as usual.

The names of buttons and boxes may differ slightly depending on the version of the app.

⑥If it does not work

When an error appears, please also read the “How to fix it” steps shown on the screen.

What you seeWhat it means, and what to do
The key is wrong or cannot be used (401)The API key is wrong or has been deleted. Redo ④ to create a new key, paste it again and press Save. When you copy it, also check that the whole key is there, from the beginning sk-or-v1- to the end, and that no extra spaces were added before or after it.
Not enough credits (402)Your credit balance is too low. Add credits at https://openrouter.ai/settings/credits. It may also stop when you reach the limit you set on the API key; in that case, raise the limit on the keys page.
The content was refused (403)Depending on the content, OpenRouter or the provider that runs the model may refuse it. Switch “LLM used for formatting” in Settings to the built-in LLM and try again. Rewording the description a little may also get it through.
Too busy (429)Too many requests were made in a short time, or OpenRouter is busy. Wait a while and try again. If you entered a name in “Provider”, emptying the box (Cheapest) lets other providers be used.
No provider found (404, 503)This appears when no provider matching the condition “only providers that do not store data” could be found at that moment (→ ⑦). Try again after a while. If you entered a name in “Provider”, check its spelling and whether that provider supports ZDR, or empty the box (Cheapest). If it keeps happening, switch to the built-in LLM.
Cannot connectCheck that you are connected to the internet. Also check that security software or your workplace network is not blocking communication with openrouter.ai.

⑦Privacy: what is sent and what is protected

When you format with OpenRouter, unlike the built-in LLM, your description and similar text are sent outside your PC. Please use it only after you understand the following.

The wording and layout of OpenRouter's pages may change without notice. If a page looks different, look for a button or item with a similar name. For fees, charges and conditions, OpenRouter's pages and terms are the correct information.

9Handy tips

⭐No need to know the tricks of the video AI (MiniMax H3): just choose or press to get the video you want (💡Examples, ✨Effects, 🎥Camera)

The video AI (MiniMax H3) has a few tricks for getting the video you have in mind: write the description in English in a fixed layout the AI reads well (the official format), point to the people in your photos in a set way, and match the video length to the spoken lines… No beginner can be expected to know all of this.

That is why WishMallet has three features with these tricks built in. Just choose or press, and a description in the form the AI handles best, plus the right settings, are filled in for you. Then simply press Create the video. Anyone can get close to the video they wanted, right away.

😵Writing everything yourself…
  • You don't know what form the AI understands
  • When the lines and the video length don't match, the speech gets slow or rushed
  • You don't know how to write “make the camera circle around”
  • Lots to decide: length, shape, detail, quality…
😊With the three features
  • Choose one and an official-format description is filled in
  • The length (matched to the lines) and the shape are set for you
  • A camera movement is one press
  • Then just press Create the video

💡Examples: choose an example and your first video is ready to go

Choose an example and the ready-made description, length, shape, detail and quality, and even the sample photo or voice, are all filled in on the Create screen. Press Create the video as is and you get a video like the example. Replace the photo with your own and it becomes your video.

✨Effects: an eye-catching effect for your own photo, just by choosing

Add a picture in ① and press an effect card. An official-format description for that effect and the length and shape that suit it are filled in. Your pictures stay as they are, so the person or pet in your photo is the star.

①Add a pictureWith “+ Add” in ①
②Choose an effect“✨Effects” → press a card
③Create the videoThe description and settings are already in

🖍️Drawing comes alive: a child's drawing starts moving on the paper

What if that drawing on the fridge could move? Add a photo of a crayon drawing in ① and press 🖍️Drawing comes alive, and the character in the drawing comes to life on the paper. The crayon strokes and the feel of the paper stay just as they are, so it looks like the child's own drawing has come alive. It also makes a lovely gift for grandparents.

A crayon drawing of a family: the sun, clouds, a house, flowers and four people holding hands
Photo addedA crayon drawing of a family
The videoThe family starts moving on the paper

Use a picture of the drawing only (do not add photos of the child who drew it).

🏪Promote your shop: a promo video from one photo and a catchphrase

Add a photo of your product or shop in ①, press 🏪Promote your shop, and type your shop's name and a catchphrase. That's all. In soft light and sparkles, the camera slowly moves in and glides around, showing off the colors and details in a vertical promo video. A cheerful narrator reads out your shop's name and catchphrase over a light, upbeat tune. It's just right for vertical videos on social media and for introducing your shop.

Inside a bakery: loaves, melon buns, sweet bean buns, croissants and more on display
Photo addedInside a bakery
Shop nameTsumugi Bakery
CatchphraseFresh-baked bread! Open every morning from 7
The videoThe camera moves in on the bread in soft light

Both are real videos made with this feature (the samples were made short and play on a loop without sound). The bakery sample was made on a Japanese screen, so its shop name and catchphrase were in Japanese.

💞AI analysis “Destined partner”: your destined partner says “Nice to meet you” from the screen

If you could meet your destined partner today, what would they look like, sound like — and what would their name be?

All you enter is your nationality, gender, birth date, name, blood type and personality. Advanced AI analysis and prediction logic, used to the full, combines four kinds of compatibility (biorhythm, name — name fortune for Japan and numerology for other countries — blood type and personality) to work out the birth date, name, blood type and personality of your best match. It even creates their face, hairstyle, build, outfit and voice, and then your one-of-a-kind destined partner looks straight at you from the video and introduces themselves. No photo is needed.

Press ▶ to play the actual videos (5 seconds each, with sound).

Real videos made with this feature on a PC set to Japanese. The greeting and subtitles follow your Windows language, in 11 languages (including English, Spanish, French, German, Chinese and Korean). Subtitles are added automatically. “Sakura Tanaka” and “Kenta Yamamoto” are fictional people created by AI. On this page, the subtitles are translated into English (the voices are in Japanese).

💘See exactly who they areName, birth date, blood type, personality and how well you match. The analysis report even explains why you two are compatible.
🎬Cast them in other videosTheir picture and voice go into the 📒Character book, so you can make a date together or a video letter from a trip.
👨‍👩‍👧With family and friendsAnyone can try it, at any age (a child's destined partner is someone their own age). Sharing the results is half the fun.
🔬What makes it special: a partner created with advanced AI analysis and prediction logic

An ordinary compatibility reading only works once you already have someone in mind (“Are we a good match?”). WishMallet's destined partner works the other way round. From your details alone, it combines four compatibility analyses — biorhythm, name, blood type and personality — with real statistics on married couples and psychology research to predict the person who matches you best. Then it brings them to life as one complete person, from their birth date, name and personality all the way to their face and voice.

An image of two glowing waves, pink and violet, gradually falling into step and turning into a heart of light
① Their birthday, worked out from real couples' statistics and biorhythms

The age gap is chosen from real statistics on the age gaps of married couples, matched to your gender and age. Then it goes through all 365 days of that year, one by one, calculating how well your three biorhythms — physical (23 days), emotional (28 days) and intellectual (33 days) — line up. It leaves out days when you both have a “critical day” (when a rhythm switches and you are said to be off balance), and makes the day you are most in tune your partner's birthday. You even see how many days in the coming year your feelings are likely to align.

An image of five glowing orbs for wood, fire, earth, metal and water, joined in a star shape above a blank antique book and a brush
② Their name: every combination of common names is scored

It pairs given names that are actually common around your partner's birth year with common surnames, and scores every single combination. Japanese names are scored with name fortune: the five grids are worked out from traditional stroke counts (Kangxi dictionary forms), and the score covers how your names' five elements get along, the fortune of the name itself and even the fortune of the married name. Names from other countries use Pythagorean numerology, matching the numbers from your birth dates and names. Only names short enough for the video's greeting are kept, so you get a believable name that's easy to say.

An image of two overlapping pentagon radar charts, pink and violet, with a heart of light in the middle
③ Personality from psychology research, blood type from compatibility charts

Your personality is measured with the Big Five, the five traits used in psychology research around the world (just pick the words that fit you, or take the 20-question detailed check). Based on research on relationship satisfaction, it then works out the partner personality you are both most likely to be happy with: kindness, dependability and emotional stability a little higher, pace of life and interests close to yours, and always within the range of real people. Allowing for uncertainty in the estimate of your own personality, it settles the answer over 500 simulations. Blood type is chosen by averaging several fortune-telling compatibility charts.

An image of a smiling woman, facing forward, taking shape from lines and particles of light
④ Face, hair, build, outfit and voice: one complete person

From the analyzed age, country and personality, it puts together their face, hairstyle, build, outfit and voice. For example, a kind person gets soft eyes and a warm voice, and a cheerful person a lively expression and a brisk way of speaking. An adult's build is based on health statistics, and the outfit even suits the season (light fabrics in summer, warm ones in winter). Finally, it gives a compatibility score that brings the four analyses together, and a confidence level that shows how reliable the result is.

😐An ordinary compatibility reading
  • Works only once you have someone in mind
  • Often uses just one method, such as star signs or blood types
  • You just read “a good match” or “not”
🤩AI destined partner
  • Predicts your best match from your details alone
  • Four analyses combined with real couples' statistics and psychology research
  • The report shows why it's them, with a confidence level
  • Creates their face and voice, so you meet them on video

The pictures are illustrations. Please enjoy the fortune part for fun.

①Open AI analysis“✨Effects” → “💞Destined partner” at the top
②Tell it about youNationality, gender, birth date (height, weight and occupation are optional) → name → blood type and personality
③See the analysis reportCompatibility, the four scores and your partner's profile
④Create the video“Use this” → “Create the video”

🖐️AI analysis “Palm reading”: have a skilled palm reader in a night alley read your palm

A night alley lit by lanterns. A skilled palm reader peers at your outstretched palm through a large magnifying glass — “Good evening. A long fire hand with short fingers.”

It feels like stepping into a real fortune teller's booth: a 15-second video from your own point of view. Making full use of advanced AI analysis, it reads your palm, works out five fortunes — love, work, money, people and vitality — and even when the turning points come, then tells you about your hand shape, your past and present, and your future.

🔬What makes it special: palm reading powered by advanced AI analysis

In an ordinary palm reading, a reader looks at your hand and reads it the way their own school taught them. WishMallet's palm reading has AI analyze the image of your palm, measure every line in numbers, and check the result against the readings of many palmistry schools.

An image of a palm with an AI scanning grid and glowing traced lines over it
① AI analyzes the image of your palm

The AI finds your palm in the photo, tells right hand from left and palm from back, and traces the lines of your palm one by one. It even picks up fine lines and branches that are easy to miss by eye, from tiny differences in brightness.

An image of a hand skeleton: points from the wrist to the fingertips joined by lines, with ruler ticks along the fingers
② Your hand shape and fingers, measured at 21 points

It finds 21 points — the wrist and the bases, joints and tips of your fingers — and measures the height-to-width ratio of your palm and the length of your fingers. Every hand is brought to the same scale before measuring, so any hand is read by the same standard, whatever its size or however the photo was taken. From this it tells which of the four hand types — earth, air, fire or water — yours is.

An image of five palm lines glowing in different colors, with antique Japanese and Western palmistry charts behind the hand
③ Checked against palmistry schools from Japan and around the world

It turns the length, slope, curve, depth, branches and breaks of your life, head, heart, fate and sun lines into numbers, and spots special lines such as a double life line or a simian line. It then checks them against the readings of many sources, from palmistry books to expert guides in Japan and around the world, and uses the readings most schools agree on. It even reads positions along your lines as ages to find when the turning points come.

😐An ordinary palm reading
  • A reader looks at your hand
  • The reading depends on the school they learned
  • You hear the result once, and that's it
🤩AI palm reading
  • AI analyzes the lines one by one and measures them in numbers
  • Checks the readings of many schools
  • Picks up fine and special lines too
  • The report and the 15-second video let you look back any time

The pictures are illustrations. Please enjoy the fortune part for fun.

Three palm readers, each with their own message and manner. Press ▶ to hear each reader's own voice (15 seconds, with sound).

All are real videos made with this feature on a PC set to Japanese. The lines and subtitles follow your Windows language, in 11 languages. The words change with your palm and the reader — a surprise until you watch. On this page, the subtitles and the words in the pictures are translated into English (the voices are in Japanese).

①Add a photo of your palm“+ Add” in ① (📋Paste, or the camera on a phone)
②Open Palm reading“✨Effects” → “🖐️Palm reading” in AI analysis → dominant hand, birth date (optional) → “Analyze”
③See the reportYour lines, the five fortunes, what each line says, timing
④Use thisThe 15-second description and settings go into the Create screen → “Create the video” or “⏳Add to queue”
A sample palm photo: the whole palm from fingertips to wrist, straight on, in good light, with the fingers slightly apart
📷 Take your palm photo like this (sample)
  • ☀️ In good light, with no shadow on your palm
  • 🖐️ Your whole palm, from fingertips to wrist, straight on
  • ✌️ Fingers slightly apart
  • 📷 On a phone, use the back camera (selfies can flip left and right)
  • Either hand works. Your dominant hand is recommended (some schools read it as your present and future).
  • If something about the photo could affect the reading, the report tells you (you can replace the photo in ① or continue as is).

📩AI analysis “Otherworld Letters”: your biggest ally in the world! A mysterious video letter that lifts your mind just by talking with a character from another world

AI looks deep into the “real feelings” behind your words and “the words you need most right now”. A character in the relationship you chose — partner, buddy and more — wraps you in a gentle world of zero lecturing and total acceptance. While you are simply having fun chatting, your heart grows lighter before you know it, and you come to like yourself a little more — a special seat reserved just for you.

What if someone in your 📒Character book actually lived in another world — and streamed live from there?

You are a viewer: you just send chat comments. The streamer (a person from your Character book) shows their face from a stream room in their own world and answers you directly, looking into the camera. But the line from the other world is very slow, so the answer arrives later as a “video letter”. When it arrives, press ✉️Reply to keep the exchange going.

Say you send just three comments, starting on a night you feel down. The streamer calls you by name, accepts how you feel, and celebrates your next-day effort as if it were their own. And by the third comment, you are the one writing, “I think I’ve changed lately.” Press ▶ to play the video letters that actually arrived (with sound).

If a woman sends them: from “Luciel”, the handsome wizardRelationship: ❤️Partner (lover / spouse) · Name to be called: “Yui”

Letter 1 YuiLuciel, listen… I made a big mistake at work today and got scolded by my boss. I really mess up everything… I cried a little on the way home.

🔬 What AI read
  • Sadness and shame (weight ●●●)
  • “I mess up everything” = an all-or-nothing habit
  • What she wants: to be accepted
→ Take her side first, before any encouragement

Letter 2 YuiYour words really worked! I got to bed early last night. Today I reported my mistake myself and turned things around. …But that’s just normal, right?

🔬 What AI read
  • Relief and pride (weight ●○○)
  • “That’s just normal” = being too hard on herself
  • What she wants: her effort recognized
→ Praise it specifically, then end with a light question

Letter 3 YuiThe most relieving moment was when I reported my mistake and my boss said, “Thank you for being honest.” And you know, I feel I’ve changed a little lately. When I fail, I no longer think “I mess up everything” like before.

🔬 What AI read
  • Relief and gratitude (weight ●○○)
  • Thinking habit: not showing ✨
  • The good change goes into the heart chart
→ Celebrate together, and say he has seen her change

📈 Heart chart (how three letters changed things)Weight of feelings ●●● → ●○○ → ●○○
Thinking habit “I mess up everything” → “that’s just normal” → gone ✨

If a man sends them: from “Alice” of WonderlandRelationship: ⚔️Buddy (bound by fate) · Name to be called: “Haruto”

Letter 1 HarutoAlice, I need advice from my buddy. Tomorrow I’m giving a presentation in front of a big crowd. Once I started thinking everyone would laugh at me and it would all be over if I failed, I couldn’t sleep.

🔬 What AI read
  • Worry and fatigue (weight ●●○)
  • Foundation: “safety” is shaky (afraid of failing, can’t sleep)
  • “They’ll laugh and it’s over” = imagining the worst
→ No “do your best” — calm him first so he can sleep

Letter 2 HarutoLast night I kept taking deep breaths and fell asleep. Thanks to that, my hands shook at today’s presentation, but I made it to the end. …It wasn’t perfect, though.

🔬 What AI read
  • Relief and pride (weight ●○○)
  • “It wasn’t perfect” = an all-or-nothing habit
  • Last tip: done ✅
→ Cheer with him and praise the step he took

Letter 3 HarutoYes, it reached them. Afterwards a junior colleague came up and said, “That was really easy to understand.” And lately presentations don’t scare me as much. Maybe next time I’ll volunteer for a big project myself.

🔬 What AI read
  • Joy and pride (weight ●○○)
  • Foundation: up to “dreams and challenges” (top of Maslow’s hierarchy) ✨
  • Thinking habit: not showing ✨
→ Tell him how dependable he’s become, and back his challenge

📈 Heart chart (how three letters changed things)Weight of feelings ●●○ → ●○○ → ●○○
Foundation (Maslow) “safety” shaky → up to “dreams and challenges” ✨

After you send a comment, the chat shows what is happening in the other world: “👀 Luciel is reading your comment” → “🔴 Luciel is answering… receiving the signal from the other world”. Waiting for the reply is part of the fun.

All are real videos made with this feature in Japanese. On this page, the subtitles are translated into English (the voices are in Japanese), and the comments are sample texts shown in English translation. In your app, the conversation lines and subtitles are spoken in your own language, in 11 languages (with English settings, the video letters arrive in English). The lines change every time with your comment and the streamer’s personality, relationship and world — a surprise until you watch. “🔬 What AI read” and the heart chart are shown only for this example; they never appear on the app screen. Luciel and Alice were added to the 📒Character book with photos from the 🎁Sample materials.

🔬What makes it special: 4 secrets that leave your heart lighter the more you talk

A typical AI chat just answers what you ask. WishMallet’s Otherworld Letters is designed to bring you the words you really need right now. All you do is enjoy chatting with a character you like. As the letters go on, your heart quietly grows lighter and you come to like yourself a little more — that is the experience we built it for.

Blue, pink and gold lights swirl like an aurora inside a speech bubble, read by a crystal ball below
1. Grasps “the words you need most right now” in an instant

“I’m just tired.” “I just want someone to listen.” — AI gently catches even the feelings you can’t put into words. Based on psychology, it analyzes 16 kinds of feelings and how much energy your heart has left, and works out what the foundation of your heart is asking for right now — “no advice, just listen” or “praise me with all you’ve got”. It also notices harsh ways of talking to yourself and the strengths that show in your words.

Hands of light cradle a glowing heart, with four orbs of light around it: a star, a sprout, a key and a footprint
2. Zero pressure on your heart: 15 ways of staying close

When your heart is tired, it never pushes a challenge — it starts by helping you feel safe. Instead of forcing you to look on the bright side, it chooses words that let your heart rest. No advice when you just want to be heard; a full-hearted celebration for good news — it picks what fits you now from 15 ways of caring. Replies that miss what you need, or that make light of your feelings, are rewritten automatically by AI, so the answer comes at the distance that feels best to you.

A stream room: a glowing crystal ball, a teacup, an old book and a lamp on a desk in front of a round window with two moons
3. Counseling wisdom, turned into “otherworld magic”

Zero jargon, zero lecturing. No advice from above, and no “Cheer up!” that corners you. A partner holds you sweetly, a buddy makes a plan right beside you, family accepts all of you, and a junior cheers you on while looking up to you — they stay in the relationship you chose (only a mentor sounds like giving advice). Even a small everyday step becomes a fun mission such as “tonight’s spell” or “a Wonderland charm”, delivered with the character’s own expressions and gestures. Now and then an answer ends with a light question (“What was the tastiest thing today?”), so the talk keeps flowing naturally.

Three flower pots in a row — a sprout, a bud and a blooming heart-shaped flower — joined by a line of light like a path of growth
4. A “heart chart” just for you: the bond deepens the longer you talk

With every exchange, your strengths and the changes in your heart are quietly noted. They remember what was on your mind and ask “How did that go?”, and rejoice with you over small growth — “You’re doing so much better than before” — becoming the one person in the world who truly understands you. If very heavy feelings go on, the character gently suggests also talking to someone you trust or a support line. Like your exchange records, the heart chart is kept only on your device (not on the developer’s servers) and is never shown on screen.

😐A typical AI chat
  • Answers what you ask, in text
  • Often does not consider how you felt when you wrote it
  • May not remember what you said last time
🤩AI-analyzed Otherworld Letters
  • AI analyzes “the words you really need right now” before answering
  • No lecturing, no “Cheer up!” — total acceptance, in the relationship you chose
  • A heart chart that understands you better the more you talk
  • A video letter whose face and voice carry the feeling

Images are for illustration. This is entertainment with a fictional character played by AI, not a substitute for medical care or counseling. If painful feelings continue, please also use the support lines.

①Choose someoneFrom your Character book
②Send a commentUp to 400 characters
③Into the queueAutomatically
④A video letter arrivesKeep going with ✉️Reply

📞Support lines: if painful feelings continue, you don't have to carry them alone. Find A Helpline lists free, confidential support lines by country. In Japan, phone and chat support lines are listed on “Mamorouyo Kokoro” by the Ministry of Health, Labour and Welfare (in Japanese).

The length for spoken lines is taken care of too

When the video is much longer than its spoken line, MiniMax H3 may speak slowly or repeat the line in the leftover time. When the video is too short, it may rush or drop part of the line. 💡Examples and ✨Effects set the video length to match how long the line takes to say (for effects where you type words, it matches the words you typed).

🎥Camera: film-like camera work without knowing the jargon

Just press a button under the description box and a camera-movement instruction goes into the description. You don't need to know filming terms such as “drone shot” or “orbit”, or how to write them for the AI. A pressed button changes color; press it again to remove the movement.

🚁From the skySlowly comes down from high in the sky
🔄Orbit aroundSlowly circles around the subject
🔍Move inSlowly moves in from far away to the subject
↔ Slide sidewaysSlides slowly sideways alongside the subject
📹HandheldHandheld, with a slight natural shake

Combine them to get even closer to what you want

①Add a pictureWith 💡Examples, the picture comes too
②Choose an effectOr 💡Examples. The description and settings go in
③Add a camera movementPress the one you like (optional)
④Create the videoThen just wait for it to finish
If it is not quite right: keep the same content, set the seed to Random and press ⏳Add to queue a few times to get several videos with different movements to compare. To adjust a finished video a little, “History” also has 🔧Adjust (Smiling, Slower and so on) and 🎨Remix the look (night, watercolor and so on) (→ 5. Using History). 🎁All three features can be used during the free trial too.

⭐Turn a character idea into a three-view sheet in moments, and give your stories a lead (🆕New character)

Just describe the character you want. The AI draws it from three angles, front, side and back (a three-view sheet), and adds it to 📒Characters automatically, together with a greeting (“Nice to meet you.”) in a voice that suits the character. Live action, anime, 3D CG, figure style… people or animals, whatever you like.

All of these were made with this feature (the first frames of the videos, as they are).

🎨Any style, any kindLive action, anime, cartoon, 3D CG, figure or mascot. Human, animal, creature or robot. Pick from the lists, or leave it on Auto.
📒Added to Characters for youThe first frame of the video (the three-view sheet) and the voice go into the Character book. The name and note are the automatic title. No need to save it yourself.
🎬Characters that stay the sameUse it from the Character book and the sheet (and voice) goes into ①. Your character keeps the same look from any camera angle.

How to use it: write and press (3 steps)

1. Pick the example💡Examples on the Create tab → 🆕New character
2. Describe your characterAny language is fine. Press Use this
3. Create the videoA short greeting video is made, and the sheet and voice go into the Character book

Make a story video with your sheet

Put the sheet in ①Choose it in 📒Characters and press Use
Write the story in ②Example: “She runs through Shibuya at dusk”
Create the videoThe same character from any angle

Characters you make stay in 📒Characters, so you can call them up quickly next time.

💡Tips
  • Make a few and keep your favorite: press ⏳Add to queue a few times to make several different sheets in one go. Each one goes into the Character book: delete the ones you don't need with 🗑Delete, and rename the ones you keep with ✏️Edit.
  • Sometimes there is no voice: if the AI could not make the voice, only the picture is saved. Create it again if you want the voice too.
  • Not the style you expected? Pick the style from the list and create it again.
  • 🪄Auto prompt is turned off for it (rewriting would break the three-view instructions).

⭐Queue them up and let it run: make several videos while you sleep (⏳Add to queue)

Add what you want to make with ⏳Add to queue, one after another, then just press ▶Start. The videos are made one by one, automatically. You don't need to sit in front of the screen until they are done.

🌙Before bed or before you go outQueue them and the rest is automatic. They are all done by the time you are back.
🎲Several takes of the same ideaWith the seed on Random, each press gives a different video. Compare them in “🎞️History” and pick the best.
🧩Many ideas at onceChange the description, photos or length and keep adding (up to 100).
1. Add to the queueSet it up on “▶️Create” and press ⏳Add to queue (as many times as you like)
2. StartCheck the list in “⏳Queue” and press ▶Start
3. Compare the resultsPick your favorite in “🎞️History”

The queue is made while the app is open (closing the app clears it). When you queue videos before bed, make sure your PC does not go to sleep. For details → ⏳Making several videos in a row (queue)

⭐Get a prompt for the video you want from just a few words (🪄Auto prompt)

Write the video you want to see in a few words and press 🪄Auto prompt. The AI rewrites it into a prompt (description) in the officially recommended format for making that video. No photos or voices are needed.

The AI plans scenes that fit the length and shape you set and fills out the description. Reading the result shows you how a description should be written to get the video you have in mind. Once you are used to it, add places, people, movements and lines, or add photos and voices, to get closer to exactly what you imagine.

Recommended steps

  1. Set the conditions first. To start, a very short video is best: Detail 0.5 / Length 1 sec / Shape Landscape 16:9 / Quality 20. It finishes quickly, so you can soon see whether your words become the video you imagined.
  2. In “②Video description”, write the video you want in a few words and press 🪄Auto prompt (with the built-in LLM, 30 seconds to 2 minutes; the first time takes a few minutes longer because the needed parts are downloaded).
  3. When the prompt is ready, press Create the video.

If formatting takes a long time, you can switch “LLM used for formatting” to OpenRouter in Settings to finish in much less time (it needs a sign-up and a small fee; → Format with OpenRouter).

For example, you can start from words like these.

Tip: do not make one sentence too long. Keep writing short sentences separated by periods, and you are more likely to get the video you imagined (for example: A giant castle floats above the clouds. A flock of white birds flies over a beautiful girl standing at its entrance).

⭐With your own photo: “I want to look younger!”, “I want to be an anime character!”, “I want to change my gender!”

All you need is one photo and 🧭Guided writing. You do not need to write a description yourself. Start by setting the video length to 1–2 seconds so that you see the result quickly.

Common steps

  1. On the “Create” tab, press + Add → 📁Choose a file and choose one photo in which the face is clearly visible. A photo of just one person, facing the front, from the chest up is recommended. You can also try JP_Mail_50.jpg from 🎁Sample files.
  2. Press 🧭Guided writing under the ② heading.
  3. Fill in “①Characters” as described for each goal below.
  4. In “Where are they and what is happening?” under “②Scenes”, write the place and the camera direction (for example: “They are sitting on a park bench. The camera shows the face from the front.”). Type a line in “What are they saying?” to make a talking video (it can be left empty).
  5. Press ✨Format prompt, and when the description is ready, press Create the video.

“I want to look younger!”

  1. In “Age, gender and occupation?”, enter the age you want to be as a number only (for example 16), the gender (for example male) and the occupation (for example Japanese).
  2. In “Features”, a short note of the look you want gets you closer (for example high school student).
  3. Do not tick “Non-human” or “Anime”.

“I want to be an anime character!”

  1. Enter the gender and occupation as they are for the person in the photo (for example male and Japanese). Age can stay empty (if you enter it, the character is of that age).
  2. Tick “Anime”.
  3. With several people in the video, tick only the person you want as anime. The other people and the background stay as in the photos, and only that person becomes anime. Tick everyone and the whole video becomes anime.

“I want to change my gender!”

  1. In Gender, enter the gender you want (for example female).
  2. Enter a number in Age too (your current age is fine). If Age is empty, the video is made with the look in the photo.
  3. Write the hairstyle and outfit in “Features” and “What are they wearing?” to get closer to that look (for example long straight hair and high school uniform).
About the resulting face: a younger face or a face of another gender is a newly created face based on features in the photo such as the eyes, nose, mouth and outline. It does not reproduce the face you really had when you were young. Think of it as “a different person, of another age or gender, who resembles you”.

Other things you can do

What you wantWhat to enter in “①Characters”
👶Become a babyAge “0”, plus Gender and Occupation
👴See yourself at 100Age “100”, plus Gender and Occupation
🌏Change the look (appearance)Age (for example “16”), Gender, and a word that describes the look in Occupation (for example “Caucasian” or “Black”)
🐈Become a catTick “Non-human”, Occupation “cat”, plus Gender (and for example “calico kitten” in Features)
🐱Become an anime catTick both “Non-human” and “Anime”, Occupation “cat”

The same list is also in 3. How to create a video, “Transform the person in one photo”.

When the change does not work well: (1) Check that a number is entered in the Age box (it is needed when you change the age, gender or look). (2) Add concrete details of the look you want to “Features”. (3) Set the seed to “Random” and create the video a few times.

When you use a photo of someone other than yourself, always get that person's permission (→ 11. Using the app safely).

⭐Make a “motion comic” from a comic page

From a single comic page, you can make a vertical short video where only the panel being read moves and talks (a motion comic). It is a popular format that can go viral on TikTok, YouTube Shorts and Instagram Reels.

Even if you ask the AI (MiniMax H3) to “move only this panel”, other panels may move too when the same character's face appears several times on the page. Lay mask images, with only the panel to move made transparent, over the video in 🎬Edit video, and the panels move one at a time.

⭐I cannot write a good description

Press 🧭Guided writing under the “②Video description” heading on the “Create” screen. Just answer the questions and the description is written for you.

Tip: “Where are they and what is happening?” works reliably when you include the place, the time of day and the movement, one of each (for example: making coffee in a café in the evening).

⭐I want to keep “one high-quality photo” without spending much time

Create a short 1-second video at good quality and keep one frame of it as a photo. It takes only a few minutes.

  1. Create a video with these settings: just one photo (no voices or reference videos) / Detail 2 (1888x1088px: Full HD) / length 1 second / any shape you like / quality 20 or more.
  2. Open 🖼Thumbnails in “History”, drag the strip at the bottom to find a good moment, and press ⭐Save to album.
  3. Open the photo on the “Album” tab, press 💾Save full quality and pick a purpose (not sure? “⭐Master image”). The AI enlarges it to just the right size and saves it as a PNG (details in 6. Best shots). 🎁During the free trial you can see the list of purposes and the preview.
Key point: normally a larger detail makes it easier to run out of memory, but with a length of 1 second and one photo, even a PC with 12 GB of video memory can create at Detail 2 (1888x1088px: Full HD). At quality 10 the picture may break down, so set the quality to 20 or more.

⭐The video does not look good

If it came out almost black or barely moving, use 🔁Make it again in “History” to bring it back to the “Create” screen, set the seed to Random and create it again. To make a few versions of the same idea and pick one, set the seed to Random and press ⏳Add to queue a few times.

  1. First, set the quality to Best and create the video again.
  2. If that is still not enough, raise the detail as far as you can. With a length of 1 second, even a low-performance PC can try up to Detail 2 (1888x1088px: Full HD).
  3. If it stops for lack of memory, buying a new graphics card is one option. But it is an expensive purchase, so before that, please also see “⭐Try the cloud without spending money” below and 7. Export, “How do cloud graphics cards compare with home ones?”. Even if you upgrade to the top home card, the full set of parts does not fit.
    💡On the other hand, if you create a lot, buying a GPU can cost less than paying for the cloud again and again → Could an expensive GPU cost less?

⭐I want to re-create a video I like with a small change

You like the framing and motion, but you want to fix only the outfit or the spoken line, or raise only the quality. In such cases, use the “seed”.

  1. On the “History” tab, press 🔁Make it again for that video.
  2. The settings come back to the “Create” screen, and the seed is set to “Same as last time”.
  3. Change only what you want to fix (the description, the quality and so on) and press Create the video.

You will not get exactly the same video. Think of it as getting close to a similar feel. Conversely, when you want to see various ideas from the same description, leave it on “Random”.

To change only the look (night, winter, watercolor and more), use 🎨Remix the look; to continue the story, use ▶Make the next part (both are in “History”).

⭐Try the cloud without spending money

Write out the same settings with 📤Export and re-create the video on a cloud service (→ 7. Export). You can start both without spending money. See “How do cloud graphics cards compare with home ones?” for how powerful a graphics card you can rent.

ServiceHow to try it without spending money
Google ColabYou can start for free. With a Google account, you need neither a sign-up nor a credit card. However, G4 and A100, which suit video creation, are GPUs you can select on a paid plan such as Colab Pro. The free tier usually offers only the T4, so in that case use a paid plan or Floyo below.
FloyoYou can use the code KATORI30 to try it free for 30 days. Create an account, choose a plan under Pricing, and enter the code in the Promo Code box on the Checkout (payment) screen. The box does not appear on the screens before you choose a plan.

The Floyo code is the one published on Floyo's official site as of September 2026. It may end at the distributor's discretion.

10Troubleshooting

Look for the heading that matches. On every screen, “How to fix it” with numbered steps appears under an error; start by following those.

🩺“Check the video engine” does not reach “Ready”

For each item shown on the screen, the way to fix it appears as numbered steps. Work through them from the top. If something is missing, press 🛠Run setup in the same “Diagnostics” section; it sets up only what is missing. When it finishes, press Check the video engine again.

🚫“Create the video” cannot be pressed

🧠Videos fail for lack of memory

  1. Close other apps that use the graphics card (browsers, games, video players and so on).
  2. On the failure screen, press Try again with detail … (lowers the detail) or Try again at … seconds (shortens the video). The “…” is the value the app proposes.
  3. Reduce the number of photos. With a reference video, use about half your usual detail or length (a reference video needs about twice the memory).

🖥“The AI could not run because the GPU is not visible” appears

The video engine (Linux in WSL) on this PC cannot use the NVIDIA GPU. This always happens inside a virtual machine, and on a physical PC it can happen right after an NVIDIA driver update or when the driver has a problem.

  1. Restart your PC.
  2. Reinstall the NVIDIA driver (580 or later) from NVIDIA's website, then restart your PC.
  3. If you are using a virtual machine (such as Hyper-V), use Windows 11 on a physical PC instead.

🧩“Guided writing” or “Auto prompt” stops with an error about a missing component

  1. Press 🛠Run setup under “Settings” → “Diagnostics” (the 🛠Run setup button shown with the error does the same). A black window opens.
  2. When it asks “Check and repair now?”, type y and press Enter (the AI models are not downloaded again; it takes about 5-15 minutes).
  3. When it finishes, reopen the app and try again.

If the black window does not open when you press the button, you can do the same thing by right-clicking setup_sglang_env.ps1 in the setup folder inside the app's install folder and choosing “Run with PowerShell”. Setup cannot start while a video is being created, a description is being formatted or a picture is being saved in full quality; wait until that finishes.

Even if this feature is unavailable, it does not affect creating videos. If you are in a hurry, switch “LLM used for formatting” in Settings to OpenRouter and you can carry on (to set it up, see Format with OpenRouter).

🌐An error appears when formatting with OpenRouter

What to do for each message (wrong key, not enough credits, content refused, too busy and so on) is summarized in Format with OpenRouter, “⑥If it does not work”. If you are in a hurry, switch “LLM used for formatting” in Settings to the built-in LLM and you can carry on.

📝“The description is too long to translate and convert into the MiniMax H3 official format here.” appears

It exceeds what the formatting AI in this app can handle at once. Repeating the same long description in every scene reaches the limit quickly.

  1. Shorten the description, or write the parts repeated in every scene only once, and try again.
  2. To use it as it is, press 📋How to convert it with ChatGPT or Gemini under the error, copy the text that appears with 📋Copy and paste it into ChatGPT or Gemini. Paste the resulting text into the description box, then press Create the video.

⚡“Create the video” cannot be pressed even though Turbo is on

If the fast model for Turbo is not on this PC yet (for example, a PC with an older setup), about 19.5 GB is still being downloaded. The progress is shown next to “Turbo”; please wait until it finishes.

🌐“The remaining free videos could not be checked” appears

During the free trial, the number of videos is counted on our server. Check that you are connected to the internet and press Check again.

🔄The update fails

Check that you are connected to the internet, wait a while, and then try “Settings” → “Updates” → Check now.

📤An exported file gives an error in another environment

How to fix Google Colab when it stops, and what we would like you to send us, are summarized in 7. Export, “When Google Colab stops or shows an error”.

💾“Save full quality” fails

  1. Close other apps that use the graphics card, such as games or video playback.
  2. Press Stop now under “Settings” → “⏸Auto shutdown of the video engine”, then try again.
  3. If it still fails, update the NVIDIA driver.

If it says “The files for AI enlargement are missing.”, reinstall the app (run the installer again). Until then, you can still save a PNG with the “📷Original size” purpose.

📮If it is still not solved

Under “Settings” → “Support”, write what the problem is and press Save support information. Attach the ZIP file it makes and contact the support address written in your purchase email.

11Using the app safely

When you prepare photos, voices and videos

License and disclaimer

12Glossary in plain words

TermMeaning
Description (prompt)Text that describes in words the video you want to create. You write it in the “②Video description” box.
Reference pictureA photo or illustration of the person you want in the video. The face and hairstyle are carried over from it.
Reference videoA video used as a reference for motion and scenery. With one, about twice the memory and time are needed.
Guided writingA screen that writes the description for you when you answer questions. Open it with the 🧭 button under the ② heading.
Auto promptA button (🪄) that rewrites what you wrote (even a few words) into an English description that suits video generation.
Built-in LLMThe AI that runs inside this PC and polishes your description. It is the default. Your text never leaves the PC.
OpenRouterAn internet service that lets you use many AIs through one door. If you choose it in Settings, formatting of the description is done by its AI (DeepSeek V4.1 Flash 1017) (→ Format with OpenRouter).
API keyA long string, like a password, that lets the app use an outside service. Do not show it to anyone.
DetailHow fine the picture is. Higher looks better but uses more video memory.
QualityHow carefully the picture is finished. Higher is more careful and takes longer.
TurboA fast finish for previews. Fine details become rougher.
SeedThe random number the picture starts from. Keep it the same to stay close to the previous framing and motion; change it and you tend to get a different result every time.
SubtitlesText shown at the bottom of the screen. It can be the same as the spoken words or different text (a translation or narration).
Graphics card (GPU) / video memory (VRAM)The part that calculates the picture, and its dedicated memory. They determine the size, length and speed of the videos you can create.
Video engineThe mechanism that actually calculates the video. It runs on Linux (WSL2) inside your PC. It stops automatically while you are not using it.
ComfyUIAn app for advanced users in which you build the way a picture is made by connecting boxes.
Google ColabA service that lets you borrow one of Google's computers (with a graphics card) from your browser and run calculations on it.
Activation codeThe code of letters and numbers you receive when you buy the app. One code can be registered on one PC.
Free trialSeven days of use without a code. You can create up to 3 videos a day (→ “Free trial” in 8.).

Last updated: 2026-10-29 (as of app 1.12.164)

▲Top