The first step in the process of scanlating any manga is obtaining the raws. The first step is to download the file the raws are in, then the second step is to extract the raws from this file.
Raws can often be found on piracy websites dedicated to hosting them (none are listed here to avoid link rot, but they can be found on the Internet). They are often lower quality than ripping the raws yourself, so this is not preferred. Be aware that some websites will add watermarks; since it is best practice to not keep watermarks from these sites (the only kind of acceptable watermark is an author watermark), one will either have to manually clean them or find a different source.
There are some general-purpose manga downloading tools; a notable one is HakuNeko. Alex's Raw Downloader is really simple to use for sites that it's compatible with (click "Info" on the site for a full list of compatible sites).
Websites that allow online viewing of manga will often render the images in a <canvas> element, which can then be extracted with canvas ripping, or using browser extensions like Canvas Downloader for Chrome, ImageAssistant for Chrome or Image Picka for Firefox. Be aware that resampling may occur. Some websites like Bookwalker add a near-invisible barcode at the bottom that points to your account, which can be difficult to remove.
A useful tool is this Stack Overflow answer which provides a script/bookmarklet that re-enables the ability to right-click on websites that have disabled it.
Usually possible to perform, but is almost always lower quality due to the resampling that may occur, and is labor-intensive. Basically, donât rip manga pages this way.
PDF raws are most common when obtaining self-published raws from websites such as https://booth.pm/en
Most raw PDFs are single pages with the images embedded in each page. So the next step is to extract the images that are embedded. There are two tools that are often used.
Canvas ripping involves obtaining raws by extracting image content from its already-rendered form on the website.
Advantages:
It usually can be performed with just a web browser and no external software.
It doesnât rely on DRM-cracking shenanigans.
If you know a bit of web dev these methods are very easy to develop and modify.
Disadvantages:
Might involve more manual work than some methods (though itâs still fast anyway).
The resulting files will be PNG files instead of the original JPEGs/WEBPs, and are thus generally larger (you can use our image compression tool to reduce their size a little, but it still wonât be as small).
You have to be careful to avoid resampling, as if the images are resized on the website before you download them, quality is reduced.
If you want to use canvas-ripping, here is a brief comparison of some online bookstores that I compared for this purpose:
BookWalker: The highest quality of the options I found.
cmoa: Lower quality than Bookwalker, at least in the preview.
k-manga: The preview resized the image 4px smaller than the original resolution, which really messes up screentones.
Kobo: Region-locked, so I didnât try it (itâs possible to insert fake details to get around the region lock).
BookLive: Slightly lower quality than BookWalker (at least in the preview).
For each site, we give a code snippet that scrapes from that site. The intended usage is to save the code snippet as a bookmark in your web browser, click the bookmark while looking at the site, and then scroll through the manga to download all of the pages. Alternatively, you can go open developer tools (e.g. via inspect element) and paste the snippets in the âConsoleâ tab.
Put the reader in vertical mode, make sure youâre zoomed out, and have the window be taller than it is wide. Tested on 2025-10-18 on Chromium; Firefox doesnât work directly, but you can use a GreaseMonkey script instead.
javascript:(()=>{const t=new Set,e=document.createElement("a"),a=document.querySelector(".dummy"),n=a.width,d=a.height;setInterval(()=>{for(const o of document.getElementById("canvasCluster").children){const c=o.children[1];if(!o.id.startsWith("wideScreen")||255!==c.getContext("2d").getImageData(0,0,1,1).data[3])continue;if(c.height!==a.height)throw new Error("pages not at full height");a.getContext("2d").drawImage(c,Math.ceil((c.width-n)/2),0,n,d,0,0,n,d);const i=o.id.slice(10).padStart(3,"0");t.has(i)||(e.href=a.toDataURL(),e.download=`${i}.png`,e.click(),t.add(i))}},100)})();
Source code
const done = new Set();
const a = document.createElement("a");
const dummy = document.querySelector(".dummy");
const w = dummy.width;
const h = dummy.height;
setInterval((() => {
for (const child of document.getElementById("canvasCluster").children) {
const canvas = child.children[1];
if (!child.id.startsWith("wideScreen") || 255 !== canvas.getContext("2d").getImageData(0, 0, 1, 1).data[3]) {
continue;
}
if (canvas.height !== dummy.height) {
throw new Error("pages not at full height");
}
dummy.getContext("2d").drawImage(canvas, Math.ceil((canvas.width - w) / 2), 0, w, h, 0, 0, w, h);
const page_number = child.id.slice(10).padStart(3, "0");
if (done.has(page_number)) continue;
a.href = dummy.toDataURL();
a.download = `${page_number}.png`;
a.click();
done.add(page_number);
}
}), 100);
GreaseMonkey userscript for Firefox
// ==UserScript==
// @name BookWalker scraper
// @version 1
// @grant GM.registerMenuCommand
// @match https://viewer-trial.bookwalker.jp/*
// @match https://viewer.bookwalker.jp/*
// ==/UserScript==
// un-override right-click
document.body.addEventListener("contextmenu", e => e.stopPropagation());
GM.registerMenuCommand("Start scraping", () => {
const done = new Set();
const a = document.createElement("a");
const dummy = document.createElement("canvas");
const w = parseInt(prompt("Enter the width in pixels of each page"));
if (isNaN(w)) return;
dummy.width = w;
setInterval((() => {
for (const child of document.getElementById("canvasCluster").children) {
const canvas = child.children[1];
if (!child.id.startsWith("wideScreen") || 255 !== canvas.getContext("2d").getImageData(0, 0, 1, 1).data[3]) {
continue;
}
const h = canvas.height;
dummy.height = h;
dummy.getContext("2d").drawImage(canvas, Math.ceil((canvas.width - w) / 2), 0, w, h, 0, 0, w, h);
const page_number = child.id.slice(10).padStart(3, "0");
if (done.has(page_number)) continue;
a.href = dummy.toDataURL();
a.download = `${page_number}.png`;
a.click();
done.add(page_number);
}
}), 100);
});
If you encounter the âpages not at full heightâ error, make sure youâre zoomed out and in vertical mode.
When using the Firefox version, youâll have to enter in the width of the pages manually. To do this, after you have zoomed out, right-click and copy (or save) any image whose content stretches from the left to the right of the page (usually the cover image), and measure its width in your preferred image editor.
If you ever want to use horizontal mode (which may require modifications to the script), keep in mind that the bottom two rows of pixels will be a barcode that identifies your user account. So make sure to crop this out.
The reason the normal bookmarklet doesnât work on Firefox is that BookWalker loads the images in an entirely different way on Firefox than Chromium â on Firefox, it uses <img> tags, while on Chromium it uses XHR. The images are always loaded cross-origin, but the <img> method allows the image to be drawn directly to a canvas without going through CORS. As a result of not using CORS, the canvas will then be considered âtaintedâ and its pixel data may not be extracted by JavaScript (unless that JavaScript is running inside GreaseMonkey).
This bookmarklet can also be used for cmoa. Put the reader in vertical mode, make sure youâre zoomed out, and have the window be taller than it is wide. Tested on 2025-08-22.
javascript:(()=>{const done=new Set,a=document.createElement("a"),cv=document.createElement("canvas"),c2=cv.getContext("2d");setInterval((()=>{for(const t of document.getElementById("content").children){if(!t.id.startsWith("content-p"))continue;const e=t.id.slice(9).padStart(3,"0");if(done.has(e))continue;const n=[...t.getElementsByTagName("img")];if(0===n.length||n.some((t=>!t.complete)))continue;const c=n.map(((t,e)=>{if(e===n.length-1)return[t,t.naturalHeight-2];const a=t.getBoundingClientRect(),c=n[e+1].getBoundingClientRect();return[t,Math.round((c.y-a.y)/a.height*t.naturalHeight)]}));cv.width=Math.max(...n.map((t=>t.naturalWidth))),cv.height=c.reduce(((t,[e,n])=>t+n),0);let o=0;for(const[t,e]of c)c2.drawImage(t,0,o),o+=e;a.href=cv.toDataURL(),a.download=`${e}.png`,a.click(),done.add(e)}}),100)})()
javascript:(()=>{const t=new Set,e=document.createElement("a"),a=document.createElement("canvas"),n=a.getContext("2d");setInterval((()=>{for(const c of document.getElementsByTagName("img")){if(!c.classList.item(0)?.endsWith("_page")||!c.alt.startsWith("page_")||!c.complete)continue;const o=c.alt.slice(5).padStart(3,"0");t.has(o)||(a.width=c.naturalWidth,a.height=c.naturalHeight,n.drawImage(c,0,0),e.href=a.toDataURL(),e.download=`${o}.png`,e.click(),t.add(o))}}),100);})()
Open developer tools (e.g. using inspect element), go to the Sources tab, and open the âPageâ sub-tab on the left.
On the left sidebar, find the file top â takecomic.jp â js/viewer â viewer.js. Right-click it and select âOverride Contentâ.
Press the {} button in the bottom-left of the middle panel to see the code.
Press Ctrl+F, and search for âdrawImageâ. There should be a section that says something along the lines of:
o.width = y * b,
o.height = w * b,
s.drawImage(c, 0, 0, r, i, 0, 0, y * b, w * b),
In these three lines, replace both instances of y * b with r, and both instances of w * b with i (of course, adjust accordingly if these variables have different names).
Save the file and reload the page.
Now use the bookmarklet:
javascript:(()=>{const a=document.createElement("a"),q=new Map;let i=0;for(const p of document.querySelectorAll("#xCVPages>.-cv-page"))q.set(p,i++);setInterval(()=>{for(const p of document.querySelectorAll(".-cv-page.mode-rendered")){if(!q.has(p))continue;a.href=p.querySelector("canvas").toDataURL(),a.download=`${q.get(p)}.png`.padStart(7,"0"),a.click(),q.delete(p)}},100)})()
This is an alternative to the canvas ripping method, but is likely less convenient and secure.
This requires downloading a modified Chromium executable on a Windows computer and the ability to run a PowerShell command.
Warning: The Chromium client downloaded here bypasses safety features and should not be used for anything other than downloading manga from Bookwalker.
These instructions apply to the latest version at the time of writing, version 0.3.3:
Bookwalker window sized for downloadingBuy the manga on Bookwalker. For most Japanese releases, you will need to use the Japanese site, not the global site.
Open a powsershell or cmd in the unzipped browser dir:
Navigate to the unzipped directory in explorer.
Shift + right-click in the folder.
Select "Open PowerShell window here".
Open the browser with command line: .\chrome.exe --user-data-dir=c:\bw-downloader-profile --no-sandbox
Log into you Bookwalker account in the launched Chromium window.
Open the book you want to rip in the viewer.
Adjust the window so exactly one page is visible.
If you can get the script to automatically click (step 6 here), that's good, but it is not working at the time of writing due to Bookwalker code changes.
Instead, you can just click through the manga you are ripping, one page at a time (giving each page enough time to load).
After you've clicked through all the pages you want to rip, the pages will be in C:\bw_export_data under a folder with a name matching the ID in the Bookwalker URL.