Steps to select imprint text and settings to create a 200×200px transparent PNG
This tool is a digital seal image generator that creates a personal-seal-style image from up to 4 characters, such as a surname. The output is fixed as a transparent 200×200px PNG, and you can adjust the imprint text, seal shape, font candidate, color, border width, and wear level on screen.
The initial state is 山田, circular shape, regular script candidate, vermilion, 3px border, and wear level 15. When using it for the first time, checking the default imprint first and replacing the surname before gradually altering shape or wear makes it easy to compare how each setting affects the visual appearance.
Three seal shapes are available: circular, vertical oval, and square. For imprint color, you can select from vermilion #E60012, red #FF0000, and dark red #A80010.
Border thickness is specified from 1px to 8px in 0.5px steps, and wear is specified from 0 to 50 in increments of 1. There is no feature allowing users to directly enter font size, character spacing, margins, or output pixel dimensions.
Because those are automatically determined by character count and fixed layout rules, it is important not to describe missing controls as customizable settings.
Changing a setting redraws the seal imprint on the browser Canvas. When wear is greater than 0, such as the default 15, pseudo-random numbers are used, so simply switching the shape or color can alter the wear pattern.
If you dislike the pattern, clicking Redraw wear generates an alternative distribution. Once finished, clicking Save as transparent PNG converts the displayed Canvas into a PNG and saves it to your device.
This operation does not directly edit Word, Excel, or PDF documents. The downloaded image must be placed separately using document software that supports image insertion.
Counting up to 4 characters without corruption and fixed 1 to 4 character layout
Input text is normalized using NFKC, stripped of leading and trailing whitespace, and restricted to a maximum of 4 graphemes. Whitespace left between characters is not automatically stripped and counts as one character unit.
In supported browsers, Intl.Segmenter splits characters into clusters close to visual units on screen, covering Japanese supplementary kanji and combined characters.
For example, 𠮷野 is treated as two characters, 𠮷 and 野, rather than being sliced midway based purely on UTF-16 code units.
In environments without Intl.Segmenter, it falls back to code point units via Array.from, so counting of combining characters may differ from supported browsers.
If 5 or more characters are entered, only the first 4 characters are retained. If the input is empty or contains only whitespace, the Canvas is cleared, redrawing and PNG saving are disabled, and a notice stating Please enter imprint text. is shown.
This control prevents accidentally saving an empty frame without text. The 4-character limit applies not only to the filename, but also to the characters actually rendered onto the Canvas.
Character layout follows fixed rules specific to this tool. A single character is placed at the Canvas center x=100, y=100 at font size 100px.
Two characters use 70px, arranged in a single vertical column at the same x=100 at upper y=66.4 and lower y=133.6. Three characters use 56px, placing the first character in the right column at x=132, y=100, and the second and third characters in the left column at x=68, y=74.8 and y=125.2.
Four characters also use 56px, assigning the first and second characters to the right column and the third and fourth to the left column.
This layout is determined automatically from the input order rather than user selection of vertical writing direction. Verify the intended order in the preview before saving.
Boundary calculation for circle, vertical oval, and square shapes, and device-dependent fonts
The outer seal border is calculated using a margin of 10px plus half of the border stroke width so that it does not clip outside the 200×200px Canvas.
For the default 3px border, the margin is 11.5px, and the circle radius is 100 - 11.5 = 88.5px. Increasing the border to 8px results in a 14px margin and an 86px radius.
These dimensions prevent the full stroke width, not just the center line, from contacting the Canvas edges even when thickened.
For vertical ovals, the vertical radius matches the circle radius, while the horizontal radius is set to 78% of that value. With a 3px border, the vertical radius is 88.5px and the horizontal radius is 69.03px.
This ratio keeps the vertical radius larger than the horizontal radius without exceeding the top and bottom of the Canvas. The square shape applies the same margin to all four sides, drawing a 177px square starting at x=11.5, y=11.5 for a 3px border.
Regardless of the chosen shape, the final output image remains a 200×200px square.
Four font options are provided: regular script candidates, semi-cursive script candidates, Mincho, and Gothic, but proprietary font files are not embedded into the image.
Regular script specifies candidates like HGKaishotai, Yu Mincho, and Hiragino Mincho ProN in order of preference, while semi-cursive specifies candidates such as Hiragino Gyosho and HGGyoshotai.
If the device lacks the primary candidate, it moves to the next, falling back to generic serif or system fonts if none are found.
Consequently, the same 山田 setting may show different glyph shapes, stroke weights, or fitting across Windows, macOS, and mobile devices.
It cannot be claimed that old seal script is selectable or that identical fonts display across all devices. Check the preview and saved PNG on your target device before submission.
Mechanism of wear levels 0 to 50 and changing patterns via redrawing
After drawing the outer border and glyphs, if the wear level is 1 or higher, Canvas ImageData is retrieved to inspect colored pixels sequentially.
Adding a Math.random pseudo-random value to a sine wave derived from each pixel's x and y coordinates produces a noise value compared against thresholds based on the chosen level.
Pixels with noise values below 1.5 times the wear level have their alpha set to 0, making them completely transparent, while values between 1.5 and 2.2 times reduce alpha to 40% of its original value, rendering them semi-transparent.
Other pixels retain their drawn opacity.
Rather than applying uniform opacity across the entire image, this process selectively chips away colored pixels that form the border and text.
Higher levels increase the chipped areas, but a specific level does not correspond to physical inkpad quantity or stamping pressure.
It is a relative 0 to 50 visual adjustment setting. Level 0 skips pixel-chipping post-processing entirely, producing a crisp imprint when shape, text, font, color, and border settings are identical.
Because Math.random generates different values on each draw, the chipped positions will not be identical on consecutive renders even with the same 山田, circle, vermilion, 3px border, and wear level 15.
Redraw wear does not reset your settings; it tests a different wear pattern while keeping current parameters intact. There is no seed parameter or history storage to reproduce an exact pattern later.
Save the PNG when satisfied, and rename files manually for specific purposes.
Safe download filenames and limitations of placing 200×200px images in documents
The saved file follows the format stamp_text.png using the imprint text. After NFKC normalization and truncation to 4 characters, characters that frequently cause issues in filenames—such as angle brackets, colons, quotes, slashes, backslashes, vertical bars, question marks, asterisks, and control characters—are replaced with underscores.
Trailing periods and spaces are removed. For example, 山/田 becomes a filename like stamp_山_田.png. Because the save button is disabled on empty input, seals without text cannot be saved.
This sanitization applies strictly to filename generation and does not attach digital certificates or access controls to the PNG.
The output is always a raster image of 200×200 pixels. There are no inputs to specify DPI, no high-resolution upscaling routines prior to saving, and no vector outputs like SVG.
Describing the tool as supporting 300 DPI or offering automatic high-resolution printing output is inaccurate. In Word and Excel, it can be inserted as an image, and in PDF documents, it can be placed using PDF software supporting image insertion, but this tool itself does not open documents, detect signature fields, or paste images into PDFs.
When inserting into documents, maintaining a 1:1 aspect ratio avoids distorting circular or square seals. While downscaling to fit document dimensions works well, magnifying the image substantially will not add outline detail beyond the original 200px and may blur edges and glyphs through interpolation.
Increasing dimensions with an image resizing tool will not add new typographic detail. Furthermore, transparent PNG rendering, layer order, print output, and PDF transparency handling depend on document software and printer settings.
Reopen the final saved document prior to submission to verify position, aspect ratio, background, and legibility.
Network communication scope of seal generation and differences from digital signatures and identity verification
Imprint text, shape, color, border width, and wear are held in React state, and seal rendering executes on the browser Canvas. The tool includes no fetch calls transmitting these parameters to an external image-generation API, nor does it persist input content or generated images to localStorage, IndexedDB, or the cloud.
Reloading the page resets inputs to default values, and previously generated seals or editing histories are not stored.
However, not using external APIs for image generation is distinct from the entire page making zero network requests. In addition to loading HTML, CSS, and JavaScript, site-wide communications for advertising and analytics may occur.
For this reason, absolute claims such as fully offline, completely transmission-free, tracking fully blocked, or eliminating all data leak risks are avoided.
Saved PNG files are stored directly on the user device, meaning location, sharing scope, backups, and deletion must be managed by the user.
The electronic seal created by this tool refers simply to a seal-like image usable on screen. It does not provide signing keys, digital certificates, timestamps, identity verification, or document tamper-detection mechanisms.
Digital signatures, as defined by Japan's Digital Agency, serve to identify the creator and verify that no alterations have occurred, which differs fundamentally from pasting a PNG image.
This tool does not guarantee digital signature functions, legal compliance, legal validity of the image, or substitution for registered seals, bank seals, or official seals.
At the same time, this tool does not dictate the legal standing of documents bearing seal images. Required authentication methods vary by document type and procedure, so check recipient, counterparty, and company rules, and consult professionals for important documents.