حكم سريع: A winning prompt is a result you got lucky with once. A template card is a result you can get again on purpose. The difference is five lines written down: the prompt, الوضعية, the style, the settings, and one note about what made it work. Write those lines the day you win, test the card three times, and you will never have to remember a prompt again — build your own card, and the next generation starts from a win.

Contents:
- 1. From One Good Result to a Card You Can Reuse
- 2. What a Card Actually Holds
- 3. Step 1: Find the Winner
- 4. Step 2: Write the Card Down
- 5. Step 3: Test the Card Three Times
- 6. What It Costs
- 7. Honest Limitations
- 8. Who This Is For
- 9. Consent and Safety
- 10. FAQ
1. From One Good Result to a Card You Can Reuse
Everyone who generates has had the same experience: one result comes out better than the rest, by a margin you cannot explain, and you have no idea how to get it again. You remember the prompt roughly and the pose vaguely. You remember nothing about the settings. So you generate again, and the magic does not repeat — because it was never in the prompt alone. It was in the exact combination of prompt, وضعية, style and settings, and you wrote none of it down.
A template card is the fix for that exact failure. It is a short, written record of a combination that worked, kept in a form you can run again. A card with five lines: what you asked for, how the body was posed, what style it was in, what settings you used, and one sentence about what made it work. Five lines, written once, and the win becomes repeatable.
This is different from using the site’s own template cards. Those are someone else’s combinations, built for a general case. Your card is built for your taste, your character, your light — and it is the one you will reach for first, because it has already won for you once.
2. What a Card Actually Holds
A card is small on purpose. If it takes more than five lines to write, you will not write it, and if it takes more than a minute to read, you will not use it.
| Line | What to write | Why it matters |
|---|---|---|
| المطالبة | The exact prompt, word for word | One changed word can change the whole result |
| وضعية | Body position in one sentence: standing, turning, seated, weight on one leg | Pose is the hardest thing to remember and the first thing to lose |
| Style | The look: realistic, soft light, warm tones, film grain, high contrast | Style is what makes two identical prompts look different |
| Settings | Model, length, مقاس, seed — whatever you set | Settings are invisible in the result and impossible to guess later |
| Note | One sentence: what made this one win | The note is the memory the other four lines cannot hold |
The prompt line is the one people get wrong most often. They write the prompt as they remember it, not as they typed it — and memory edits. It drops the adjective that did the work and keeps the adjective that did nothing. Copy the prompt from the generation history, character for character. If the platform keeps no history, copy the prompt into the card before you press generate.
The note line is the one people skip, and it is the one that matters most a month later. “Warm window light, face turned three-quarters, hair loose” tells you what to protect when you vary the card. “Good one” tells you nothing. The note should name the one thing that made this result better than the version before it. If you cannot name it, the result was luck, and a card built on luck will not repeat.
3. Step 1: Find the Winner
A card starts with a win, and a win needs a definition. Not “the best one” in some absolute sense — the one that did the job you needed, better than the attempts around it. The winner is the result you would show someone else. It is the result you kept coming back to. It is the result that made the previous three attempts look like practice.

The way to find winners is to generate in sets. Run the same idea three to five times with small variations — one word changed, one pose shifted, one light moved — and compare the set side by side. The winner is usually obvious: one result has the face, the pose and the light all at once, while the others have two out of three. If no result stands out, the idea needs a bigger change, not a card. A card records a win. It cannot create one.
There is a trap here: building a card from a result you liked for the wrong reason. A result can be striking because of a flaw — a strange light, an odd angle, a face that looks intense because it is slightly off. If you card the flaw, the card will repeat the flaw. Before you write the card, ask one question: would I want this exact quality in the next ten results? If not, enjoy the result and do not card it.
4. Step 2: Write the Card Down
Writing the card takes five minutes and it should happen the day you win, while the details are still in your head. The format does not matter — a note app, a text file, a physical index card. What matters is that the five lines are all there, and that the card lives somewhere you will actually look.

Start with the prompt line, copied character for character. Then the pose, in one sentence. Then the style, in three or four words. Then the settings, exactly as you set them — model, مقاس, length, seed, anything you touched. Then the note: one sentence about what made this one win. If you are tempted to add a sixth line, the card is trying to hold two wins at once. Split it into two cards.
Name the card after the win, not after the prompt. “Warm window, three-quarter face” is a name you will recognize in a month. “Prompt v7 final FINAL” is not. The name is the first line you read when you are choosing a card, so it should tell you what the card does. And keep the cards together: one folder, one file, one deck. A collection scattered across three apps is three collections, and three collections are no collection.
5. Step 3: Test the Card Three Times
A card that has not been tested is a wish, not a tool. The test is simple: run the card three times, on three different days, and see whether the win repeats. Three runs are enough to separate a real combination from a lucky draw. If the card wins twice out of three, it is a card. If it wins once out of three, it was luck, and the card needs a rewrite — usually the prompt line, which is where luck hides.
The test also tells you what the card is sensitive to. Run it once with a different input photo, once with a different light, once with the same everything. If the card survives the new photo but fails the new light, it is light-sensitive, and the note should say so. If it fails the new photo, it is photo-sensitive. A tested card knows its own limits. An untested card discovers them at the worst possible moment — when you are counting on it.
When the card passes, it goes into the deck. The deck is the set of cards you actually use, and it should stay small: five to ten cards, each one tested, each one named after its win. A deck of thirty untested cards is a pile. A deck of seven tested cards is a toolkit. When a card stops winning — when the model changes, or your taste changes — retire it. A card that no longer wins is a souvenir.
6. What It Costs
الأسعار تم التحقق منها 13 نوفمبر 2026. Only the numbers below are used in this article; they are the platform’s published credit prices.
Writing a card costs nothing. The cost is in the testing. Five text-to-image generations to find a winner cost 30 credits — $0.60 at the $19.99 قطيع ($0.020 per credit) or $0.36 at the $59.99 قطيع ($0.012 per credit). Three test runs of the card at text-to-image prices cost 18 الأرصدة: $0.36 or $0.22. If your card is a template generation, the same three tests cost 147 الأرصدة: $2.94 or $1.76. A full cycle — five generations to find the win, three to test the card — costs 48 credits as text-to-image ($0.96 / $0.58) or 392 credits as template generations ($7.84 / $4.70).
The packs themselves: $19.99 buys 1,000 credits and $59.99 buys 5,000 الأرصدة, both one-time purchases with instant credit. The annual packs are $119.88 $359.88, paid once and delivered monthly across 12 months. There is no auto-renewal wording on these packs, and no $199 pack exists yet. An undress costs 21 الأرصدة; a face swap costs 26; a 5-second video template costs 351 and an 8-second one costs 561.
7. Honest Limitations
A card cannot beat the model. A card records a combination that worked once under one set of conditions. It does not guarantee the combination will work again, and it certainly does not guarantee it will work on a different input, a different model, or a different day. Three successful tests make a card reliable, not certain.
There is also a limit to how many cards help. Five to ten tested cards cover almost every mood and every job. Past ten, the deck becomes a second library to maintain, and the time you spend choosing a card starts to cost more than the time the card saves. The point of a card is to skip the thinking — a deck so large that choosing takes longer than generating has defeated its own purpose.
أخيراً, cards go stale. Models change, and a prompt that won in November may not win in January. A card that has not been run in a month should be re-tested before you count on it, and a card that fails twice in a row should be retired.
8. Who This Is For
This article is for you if you generate often enough to have favorites — if you have ever thought “I wish I could get that one again” and meant it. It is for you if you have a character, a look, or a mood you return to, and you are tired of retyping the same prompt from memory and getting a different result every time. A card is a small investment that pays off every time you use it.
Skip it if you generate once in a while and every result is a new idea. A card is for repetition, and if nothing repeats, there is nothing to card. Skip it, too, if you are happy with the site’s own template cards: they are tested, they are maintained, and they cost nothing to use. Your own card is for the win the site’s cards never gave you.
9. Consent and Safety
Only build cards from results of yourself or of fictional adults you created. Never build a card from an undress or a face swap of a real person who did not agree to it, and never build anything involving a minor — real or realistic — under any circumstances. A card makes a result repeatable, and a repeatable result travels further than a one-off. The further it travels, the more important it is that you had the right to make it in the first place.
Keep your deck private. A card that records a prompt, a pose and a style is a recipe for a specific kind of image, and a recipe shared carelessly is a recipe used carelessly. And keep your source photos private: the platform deletes uploads 24 hours after upload once all referencing tasks have finished, but anything you generate from a card is yours to answer for once it leaves your account.
10. FAQ
What is a template card, exactly?
A short written record of a combination that worked: the exact prompt, the pose in one sentence, the style in a few words, the settings you used, and one sentence about what made it win. Five lines, written once, and the win becomes repeatable.
How do I know a result is worth carding?
Ask one question: would I want this exact quality in the next ten results? If yes, card it. If the result is striking because of a flaw, enjoy it and do not card it. Cards are for qualities you want to repeat.
Why must I copy the prompt exactly?
Because memory edits: it drops the adjective that did the work and keeps the adjective that did nothing. Copy the prompt from the generation history, character for character. If the platform keeps no history, copy the prompt into the card before you press generate.
How many times should I test a card?
Three times, on three different days. If the card wins twice out of three, it is a card. If it wins once out of three, it was luck, and the card needs a rewrite — usually the prompt line. The test also tells you what the card is sensitive to: a new photo, a new light, or neither.
How many cards should my deck hold?
Five to ten, each one tested, each one named after its win. A deck of thirty untested cards is a pile. When a card stops winning, retire it. A card that no longer wins is a souvenir.
Can I share my cards with other people?
You can, but keep the deck private by default. A card is a recipe for a specific kind of image, and a recipe shared carelessly is a recipe used carelessly. Never build or share a card from a result of a real person without their agreement, and never build anything involving a minor.
تابع القراءة:
- Outgrowing the Card: When to Stop Using Templates and Generate From Scratch
- Your First 10 Template Cards: A Starter Set for Every Mood
- The Template Doesn’t Fit: Fix the Mismatch First
Template cards and tools live on the AI Tools صفحة; image generation begins on the مولّد الإباحية بالذكاء الاصطناعي صفحة, and prompt craft is covered in the الدليل الفوري للإباحية بالذكاء الاصطناعي. Undress results are made on the التعرية بالذكاء الاصطناعي صفحة, and face work is on the AI صورة الوجه مبادلة صفحة.