96 Pindou
← Tutorial Center96 PINDOU · LEARN BY MAKING

Make a photo easier to bead before turning it into a pattern.

Learn how AI credits are used, the difference between direct photo conversion and AI chibi conversion, and how to choose photo redrawing, text creation, eight styles, prompts, and sizes.

Hands-On Cat & Dog ExampleStep-by-step illustrated guideFor Phones and Computers

Watch and follow along

View all 9 videos ↗
01 / AI生图

Create an AI chibi image before converting to beads

Compare the same pet photo converted directly and after AI chibi styling, and learn why simplifying complex photos helps.

Chinese narration and captions00:58
Link to this video ↗

01AI credits draw the image; the chart tool turns the image into beads

AI credits are used for photo redrawing and text-to-image generation. Each successful AI image generation consumes 1 AI credit; failed generations refund the reserved credit, and failed works are not added to the AI works list. When the same reference image, prompt, and other generation conditions match one of your own successful previous results, you can reuse it for free.

Converting a regular image into a pattern does not consume AI credits. It reduces the image to a grid, finds a nearby chart color for each cell, and counts the beads. Access to the pattern tools is enabled separately. AI credits and pattern access are different benefits for different purposes.

Why do direct photo conversions sometimes have fragmented colors and unclear facial features?

A photo may contain millions of pixels, but a 104×104 grid has only 10,816 positions, and colors must also be chosen from a limited brand palette. Hair, skin texture, background, and subtle shadows are squeezed into the same cells. After resizing and limiting colors, they can become speckles, while eyes and noses may blend into surrounding colors. Direct conversion only samples and matches colors; it does not understand which lines are important outlines.

AI first organizes the subject: it simplifies facial features, reduces hair detail and background distractions, and turns fine gradients into larger color blocks. When converting to a chart, the limited grid can represent the subject more clearly. AI may also change details, so check people’s features, pets’ markings, and the number of people in group photos.

Not every image needs AI first. Clear cartoons, pixel art, icons, and text can usually be converted directly. Complex photos, small portraits, and pet memorials are better suited to chibi simplification first.

02The same cat-and-dog photo, using two conversion routes

Both fuse bead results use a 104×104 grid, up to 24 colors, and the same MARD color palette. The difference is the source: the left uses direct photo conversion, while the right first uses AI to organize the image into a chibi style. Click an image to enlarge it and compare the eyes, ears, and hair edges.

01 Original photoRich layers of fur, lighting, and background
02 Convert Directly to a PatternCompressing details into limited cells can create scattered colors
03 AI chibi redrawEmphasize facial features and simplify fur and backgrounds
04 Convert again to fuse bead chartOutlines and large color blocks are easier to recognize

These comparison images were made using this site’s actual materials and conversion features. The examples illustrate the workflow; actual results vary with the source image, size, color limit, and generated content.

03Two entry points—choose based on your creative workflow

The AI image-generation pop-up in the console

Use this while simplifying source material alongside the pattern. The existing dialog flow is preserved, and after success you can continue to pattern conversion without leaving the console.

The independent AI image-generation page in the left menu

Use this first when exploring different images. It has separate source materials, styles, and results. After generation, you can download the result or explicitly click “Make fuse bead chart.”

Redraw Photo: Preserve a Specific Person, Pet, or Scene

Choose “Photo Redraw” and upload a JPG, PNG, or WEBP file, up to 12MB per image. The subject should be clear; avoid images that are too dark, heavily obstructed, or show people too small. In group photos, everyone’s face should be visible. For pet photos, choose images where the eyes, ears, and main markings are clearly visible.

Text Creation: Start with an Idea

Choose “Text Creation” if you do not need a reference photo, but describe the full scene. For example: “An orange-and-white cat and a cream-colored curly-haired puppy sitting side by side, facing forward, full bodies, big eyes, white background, clear large color blocks, no text.” Compared with writing only “make it cuter,” describing the subjects, poses, background, and details makes it easier to express what you want.

How to write “What small details do you want to keep?”

In Photo Mode, add essential features that must be preserved, such as orange-and-white markings, folded ears, glasses, scarves, and clothing colors. Do not pack too many conflicting requirements into 160 characters; preserve the two or three most important features first, then reduce the background and decorations. When processing the selected photo and description, the AI uses the service provider configured by this site.

04Eight styles and finished-size options

StyleWhat Is It Suitable For?Recommended
AI chibi artPortraits and everyday memories; recommended by default.Large eyes and small proportions—check whether the character’s features are preserved.
Pet portraitCat and dog portraits, pet memories.Keep the main markings and ear shapes while reducing fine fur detail.
Couples & groupsCouples, friends, parent and child, or photos with pets.Check that the number of people, their positions, and facial features are complete.
Mini charmKeychains, small gifts, and small patterns.Reduce complex accessories and prioritize the subject’s outline.
Full-body dollCommemorate outfits, poses, and movements.The original photo should include the full body; check that hands and feet have not been cropped.
Pixel illustrationRetro characters and game styles.Emphasizes clear color blocks; you still need to match the bead brand’s color chart afterward.
Cute stickerRounded icons and gift decorations.Emphasize outlines and check whether fine lines remain connected at small sizes.
Minimal silhouetteFor beginners, small designs, and projects with fewer colors.Accept that some details will be simplified, prioritizing a recognizable main subject.

Free Size or 52×52 Cells?

A free size lets AI preserve more details. After generation, you can choose the actual grid size in the console. 52×52 cells guides AI toward a simpler composition suitable for small projects; it does not mean the AI output is already an exact 52×52 pattern using the brand color codes. The final rows, columns, colors, and bead count are determined by the pattern conversion tool.

When making a small keychain, choose a simple style and small target size first instead of shrinking a complex image afterward. Complex details cannot all fit into a small number of cells.

05Check the result first, then decide what to do next

  1. Log in, check your AI credit balance, choose a photo or write a description, then click Generate.
  2. Wait for processing to finish. Do not submit repeatedly while waiting. If it fails, the reserved AI credits will be returned, and you can try again later.
  3. First check whether the subject, number of people, facial features, and background look right. Successful results are saved to the current account’s AI works.
  4. The independent page lets you “Download Original Image” directly. Click “Make Fuse Bead Pattern” to continue to conversion. The console pop-up continues the existing image-to-pattern workflow.
  5. Choose the brand, grid size, and color limit in the dashboard, then check the cats’ and dogs’ eyes and outlines. Use the editor to modify cells individually if needed.

You can find successful images later under “AI Works,” download them, or continue making charts. Confirm that you have saved any images you need before deleting a work. Failed results are not retained as works.