<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: disankyo</title>
    <description>The latest articles on DEV Community by disankyo (@disankyo).</description>
    <link>https://dev.to/disankyo</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4174464%2Ff03ed2fc-ac2d-4432-ac5f-3bfec317d01d.jpg</url>
      <title>DEV Community: disankyo</title>
      <link>https://dev.to/disankyo</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/disankyo"/>
    <language>en</language>
    <item>
      <title>A Pre-OCR Checklist for Scanned Documents</title>
      <dc:creator>disankyo</dc:creator>
      <pubDate>Tue, 17 Nov 2026 02:27:56 +0000</pubDate>
      <link>https://dev.to/disankyo/a-pre-ocr-checklist-for-scanned-documents-1260</link>
      <guid>https://dev.to/disankyo/a-pre-ocr-checklist-for-scanned-documents-1260</guid>
      <description>&lt;p&gt;OCR accuracy is often determined before the recognizer even runs. Five minutes of scan preparation can save an hour of tweaking OCR settings.&lt;/p&gt;

&lt;p&gt;Here's a quick checklist:&lt;/p&gt;

&lt;ol&gt;
&lt;li&gt;
&lt;strong&gt;Deskew the pages first.&lt;/strong&gt; Even a slight rotation can throw off column detection, especially in tables.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Crop out scanner borders.&lt;/strong&gt; Dark edges can be mistaken for characters or table boundaries.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Separate photos from text-heavy pages.&lt;/strong&gt; Mixing the two often leads to inconsistent recognition results.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Check page order and orientation before running a batch.&lt;/strong&gt;&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Keep the original scan alongside the extracted text.&lt;/strong&gt; This gives someone a way to verify lines the OCR engine isn't confident about.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;Some pages will still be difficult to process. Handwritten notes, stamps covering text, and low-contrast carbon copies are common trouble spots. Flag them for human review rather than quietly publishing text that may be wrong.&lt;/p&gt;

&lt;p&gt;One more check if you're creating a searchable PDF: make sure the invisible text layer lines up with the visible text on each page. A misaligned layer can be worse than having no text layer at all, because searches may return the wrong page or location.&lt;/p&gt;

</description>
      <category>webdev</category>
      <category>pdf</category>
      <category>tutorial</category>
    </item>
    <item>
      <title>Why Your GIF Looks Muddy Even at the Same Bitrate</title>
      <dc:creator>disankyo</dc:creator>
      <pubDate>Sun, 15 Nov 2026 04:35:51 +0000</pubDate>
      <link>https://dev.to/disankyo/why-your-gif-looks-muddy-even-at-the-same-bitrate-3i31</link>
      <guid>https://dev.to/disankyo/why-your-gif-looks-muddy-even-at-the-same-bitrate-3i31</guid>
      <description>&lt;p&gt;Two GIFs with the same file size can look surprisingly different. The way their color palettes are built explains much of that difference.&lt;/p&gt;

&lt;p&gt;Try this order:&lt;/p&gt;

&lt;ol&gt;
&lt;li&gt;
&lt;strong&gt;Build a global palette:&lt;/strong&gt; Sample multiple frames to generate a global palette instead of relying on the first frame alone.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Decide whether to use dithering:&lt;/strong&gt; It helps preserve the appearance of gradients, but can make flat UI screenshots with small text look noisier.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Drop frames before reducing colors:&lt;/strong&gt; Fewer frames with more consistent colors often preserve the impression of motion better than keeping every frame and sacrificing color quality.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;Which step matters most depends on the source material. A screen recording with dense text needs a different palette strategy from a clip of moving scenery. No single preset works well for both.&lt;/p&gt;

&lt;p&gt;GIF is also an indexed-color format. With smooth gradients or subtle skin tones, no amount of palette tuning will make it look exactly like the original video. Use a modern format when visual fidelity matters, stick with GIF when a platform requires it, and always judge the result at the size it will actually be displayed.&lt;/p&gt;

</description>
      <category>webdev</category>
      <category>tutorial</category>
      <category>productivity</category>
    </item>
    <item>
      <title>Word to PDF Without Losing Formatting — What Actually Works</title>
      <dc:creator>disankyo</dc:creator>
      <pubDate>Sat, 14 Nov 2026 02:37:41 +0000</pubDate>
      <link>https://dev.to/disankyo/word-to-pdf-without-losing-formatting-what-actually-works-25gm</link>
      <guid>https://dev.to/disankyo/word-to-pdf-without-losing-formatting-what-actually-works-25gm</guid>
      <description>&lt;h1&gt;
  
  
  Word to PDF Without Losing Formatting: What Works
&lt;/h1&gt;

&lt;p&gt;Few things are more frustrating than a PDF with shifted margins or replaced fonts. Here are three reliable ways to preserve the original layout:&lt;/p&gt;

&lt;ol&gt;
&lt;li&gt;
&lt;strong&gt;LibreOffice Export as PDF&lt;/strong&gt; — A solid offline option for keeping the layout intact.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Microsoft Word Save as PDF&lt;/strong&gt; — Works well for most straightforward documents.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;An online converter with a proper layout engine&lt;/strong&gt; — For example, &lt;a href="https://www.master-converter.com/en/tools/word-pdf" rel="noopener noreferrer"&gt;Master Converter&lt;/a&gt;. It charges per file, with no subscription, making it useful for occasional conversions.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;A few things to watch for:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Missing embedded fonts:&lt;/strong&gt; These can be replaced during conversion. Embed fonts in the original document when possible.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Complex tables:&lt;/strong&gt; Convert a sample page first and check the layout.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Large documents:&lt;/strong&gt; If a conversion times out, try splitting the file into smaller parts.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;If the formatting still comes out wrong, try LibreOffice as a fallback. It is often a reliable choice.&lt;/p&gt;

</description>
      <category>pdf</category>
      <category>productivity</category>
      <category>tooling</category>
      <category>saas</category>
    </item>
    <item>
      <title>Alpha vs Solid Fill: Choosing a Background Removal Output</title>
      <dc:creator>disankyo</dc:creator>
      <pubDate>Fri, 13 Nov 2026 03:08:29 +0000</pubDate>
      <link>https://dev.to/disankyo/alpha-vs-solid-fill-choosing-a-background-removal-output-12</link>
      <guid>https://dev.to/disankyo/alpha-vs-solid-fill-choosing-a-background-removal-output-12</guid>
      <description>&lt;p&gt;A cutout can look perfect against a checkerboard background and still look wrong when placed on an actual page. In many cases, the problem comes down to the output format.&lt;/p&gt;

&lt;p&gt;Choose the format based on where the image will be used:&lt;/p&gt;

&lt;ol&gt;
&lt;li&gt;
&lt;strong&gt;Use a transparent PNG (with an alpha channel)&lt;/strong&gt; when the subject needs to work on different backgrounds. Preserve the soft pixels around the edges. Cutting them off too aggressively can create a visible halo.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Use WebP or JPEG with a solid background&lt;/strong&gt; when the destination is fixed, such as a product grid. You can merge the background into the image once instead of carrying around a large alpha channel.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Check the result at its actual display size.&lt;/strong&gt; Edge artifacts that are hard to notice at full resolution can become obvious fringes in a thumbnail.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;Fine hair, glass, and semi-transparent fabrics are still difficult for automated background removal tools. If the image is important, leave room for a manual touch-up rather than assuming the model has handled every detail.&lt;/p&gt;

&lt;p&gt;Color profiles are another easy detail to overlook. Even with a clean mask, compositing an image in a different color profile can make the result look slightly gray or dull.&lt;/p&gt;

</description>
      <category>ai</category>
      <category>tutorial</category>
      <category>webdev</category>
    </item>
    <item>
      <title>The Case for Keeping Document Conversion Stateless</title>
      <dc:creator>disankyo</dc:creator>
      <pubDate>Thu, 12 Nov 2026 11:10:26 +0000</pubDate>
      <link>https://dev.to/disankyo/the-case-for-keeping-document-conversion-stateless-2846</link>
      <guid>https://dev.to/disankyo/the-case-for-keeping-document-conversion-stateless-2846</guid>
      <description>&lt;p&gt;Most document tools keep files around longer than necessary. Sometimes they end up in a temp bucket, a cache, or an old debug log. Often, nobody explicitly decided to keep them—they just never got deleted.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Stateless conversion starts with a clear retention rule:&lt;/strong&gt; when a file arrives, how long it can be processed, and when it is deleted. Write that down. Deletion shouldn't be treated as an invisible background task.&lt;/p&gt;

&lt;p&gt;There is also a practical reason to keep file retention to a minimum. &lt;strong&gt;The fewer files a system keeps, the less data it has to protect, move, or account for later.&lt;/strong&gt; When a customer asks what happens to their documents, the answer is much simpler when the system doesn't keep them around in the first place.&lt;/p&gt;

&lt;p&gt;A stateless design has trade-offs. &lt;strong&gt;If a browser closes halfway through a job, there may be nothing left to resume.&lt;/strong&gt; That's not necessarily a flaw—it is simply a consequence of choosing not to retain user files. It's better to be clear about that limitation than to imply that a recovery feature exists when it doesn't.&lt;/p&gt;

</description>
      <category>saas</category>
      <category>privacy</category>
      <category>productivity</category>
      <category>documentation</category>
    </item>
    <item>
      <title>Latency Budgets for Browser-Based Voice Previews</title>
      <dc:creator>disankyo</dc:creator>
      <pubDate>Wed, 11 Nov 2026 02:53:57 +0000</pubDate>
      <link>https://dev.to/disankyo/latency-budgets-for-browser-based-voice-previews-1c47</link>
      <guid>https://dev.to/disankyo/latency-budgets-for-browser-based-voice-previews-1c47</guid>
      <description>&lt;p&gt;A voice preview that takes too long feels broken even when the audio is perfect. The fix is usually not a faster model; it is a clearer latency budget.&lt;/p&gt;

&lt;p&gt;Split the wait into three parts:&lt;/p&gt;

&lt;ol&gt;
&lt;li&gt;Time to first byte: show an explicit queued state immediately, not a spinner with no label.&lt;/li&gt;
&lt;li&gt;Time to first audio chunk: stream partial audio instead of waiting for the whole render.&lt;/li&gt;
&lt;li&gt;Time to usable result: let the user replay and re-word without re-uploading the sample.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;The trap is caching everything. A cached sample is fine, but caching a generated line means a typo in the script forces a full re-record. Cache the voice identity, not the sentence.&lt;/p&gt;

&lt;p&gt;Short samples cannot reproduce every speaking style. If a script needs heavy emphasis or code-switching, say so in the UI instead of letting the preview fail quietly.&lt;/p&gt;

&lt;p&gt;Keep a per-stage timing log. It tells you whether to optimize inference, transport, or the interface, and it stops you spending a week speeding up the wrong stage. The numbers will be different on every stack; the three-way split will not.&lt;/p&gt;

</description>
      <category>ai</category>
      <category>productivity</category>
      <category>saas</category>
      <category>design</category>
    </item>
    <item>
      <title>What We Learned Letting People Preview a Cloned Voice First</title>
      <dc:creator>disankyo</dc:creator>
      <pubDate>Tue, 10 Nov 2026 05:51:02 +0000</pubDate>
      <link>https://dev.to/disankyo/what-we-learned-letting-people-preview-a-cloned-voice-first-17o</link>
      <guid>https://dev.to/disankyo/what-we-learned-letting-people-preview-a-cloned-voice-first-17o</guid>
      <description>&lt;p&gt;The first version of our onboarding asked people to commit before they heard anything. They did not commit.&lt;/p&gt;

&lt;p&gt;So we moved the preview to the front. The change was small and the effect was not: people started testing awkward sentences on purpose, which is exactly the behaviour we wanted. A preview is not a demo, it is a debugging tool for expectations.&lt;/p&gt;

&lt;p&gt;Three things came out of it. First, users judge a voice by one difficult sentence, not by an average. A proper noun, a number, a borrowed word. Second, they want to change the script without starting over. Third, they notice a bad preview faster than they notice a slow one.&lt;/p&gt;

&lt;p&gt;What we got wrong: we assumed a longer sample would always help. It does not. A clean short recording beats a noisy long one, and pretending otherwise wastes everyone's time.&lt;/p&gt;

&lt;p&gt;We also stopped adding a disclosure line to the end of every post. It belongs where the product is actually mentioned, not as a template.&lt;/p&gt;

&lt;p&gt;Resona is pay-per-use with one-time credits, no subscription, and credits that never expire: &lt;a href="https://resonaio.com" rel="noopener noreferrer"&gt;https://resonaio.com&lt;/a&gt;&lt;/p&gt;

</description>
      <category>ai</category>
      <category>productivity</category>
      <category>voice</category>
      <category>audio</category>
    </item>
    <item>
      <title>Merging PDFs Without Re-Encoding: Where the Bytes Actually Go</title>
      <dc:creator>disankyo</dc:creator>
      <pubDate>Mon, 09 Nov 2026 05:50:39 +0000</pubDate>
      <link>https://dev.to/disankyo/merging-pdfs-without-re-encoding-where-the-bytes-actually-go-52a1</link>
      <guid>https://dev.to/disankyo/merging-pdfs-without-re-encoding-where-the-bytes-actually-go-52a1</guid>
      <description>&lt;p&gt;A common problem with PDF merge tools is that the output can end up larger or lower in quality than the original files. This often happens when pages are rendered and exported again instead of being merged at the document level.&lt;/p&gt;

&lt;p&gt;A few things worth checking:&lt;/p&gt;

&lt;ol&gt;
&lt;li&gt;Check each input before merging. Look at page count, page dimensions, embedded fonts, and whether the PDF is already optimized for web viewing.&lt;/li&gt;
&lt;li&gt;Preserve the original page content wherever possible. A PDF merge should combine page objects, not turn every page into an image and rebuild the document.&lt;/li&gt;
&lt;li&gt;Check the merged file afterward. Try selecting and copying text, inspect the embedded fonts, and compare the output with the originals to catch unexpected changes.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;There are limits, though. Different color profiles can cause visual differences, and scanned PDFs are already image-based. A merge won't fix either problem. If consistent colors or image quality matter, handle those issues before combining the files.&lt;/p&gt;

&lt;p&gt;It's also worth keeping a reference copy and comparing the output after changes to your merge process. That helps catch regressions before they affect users.&lt;/p&gt;

&lt;p&gt;We built PDF Converter with one-time credits instead of a monthly subscription. Credits never expire, so you only pay when you need to work with PDFs: &lt;a href="https://master-converter.com" rel="noopener noreferrer"&gt;https://master-converter.com&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;Disclosure: I work on PDF Converter.&lt;/p&gt;

</description>
      <category>productivity</category>
      <category>saas</category>
      <category>tooling</category>
      <category>webdev</category>
    </item>
    <item>
      <title>Video to GIF Without Losing Quality</title>
      <dc:creator>disankyo</dc:creator>
      <pubDate>Sat, 07 Nov 2026 05:14:03 +0000</pubDate>
      <link>https://dev.to/disankyo/video-to-gif-without-losing-quality-42jj</link>
      <guid>https://dev.to/disankyo/video-to-gif-without-losing-quality-42jj</guid>
      <description>&lt;p&gt;A GIF is the easiest way to drop a clip into a README, a chat, or a tweet — until it lands at 14 MB and pixelated. The two knobs that decide whether your GIF looks good &lt;em&gt;and&lt;/em&gt; stays small are &lt;strong&gt;frame rate&lt;/strong&gt; and &lt;strong&gt;size&lt;/strong&gt;, and they pull in opposite directions. Get the balance wrong and you either get a slideshow or a file too heavy to embed.&lt;/p&gt;

&lt;p&gt;Here is how to think about it, and a workflow that lets you tune both before you commit a single credit.&lt;/p&gt;

&lt;h2&gt;
  
  
  The trade-off in one sentence
&lt;/h2&gt;

&lt;p&gt;Lower the frame rate to shrink the file; lower the dimensions to shrink it more. Quality loss comes from dropping frames (motion looks choppy) and from scaling down (detail disappears). Your job is to spend those trade-offs where they hurt least.&lt;/p&gt;

&lt;h2&gt;
  
  
  Step 1 — Start from the right source
&lt;/h2&gt;

&lt;p&gt;GIF encodes &lt;em&gt;every frame as a full image&lt;/em&gt;, which is why it balloons. Two things you can control at the source:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Trim first.&lt;/strong&gt; A 3-second loop beats a 12-second clip every time. Cut the clip to the exact moment with a trim tool, then convert. You are literally removing 75% of the frames before GIF even sees them.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Crop to the action.&lt;/strong&gt; If the relevant part is a 480×270 corner of a 1920×1080 screen recording, crop it. Every pixel you drop is a pixel GIF does not have to store across hundreds of frames.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Step 2 — Pick the frame rate on purpose
&lt;/h2&gt;

&lt;p&gt;This is the lever most people leave on autopilot.&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;UI, code, static scenes:&lt;/strong&gt; 10–12 fps is plenty. Nothing moves fast, so the eye fills in the gaps.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Talking head / gameplay:&lt;/strong&gt; 15–20 fps keeps motion smooth without going wild on size.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Fast action (sports, particles):&lt;/strong&gt; 24–25 fps, but expect the file to grow fast — each added frame is a full image.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Rule of thumb: drop the frame rate to the lowest value where the motion still reads, then stop. 15 fps is the sweet spot for most clips people actually share.&lt;/p&gt;

&lt;h2&gt;
  
  
  Step 3 — Cap the size, not just the rate
&lt;/h2&gt;

&lt;p&gt;Frame rate alone will not save you if the clip is huge. Pair it with a dimension cap:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Target &lt;strong&gt;480–640 px on the long edge&lt;/strong&gt; for chat and social embeds.&lt;/li&gt;
&lt;li&gt;Keep &lt;strong&gt;under ~6 seconds&lt;/strong&gt; for anything meant to loop in a feed.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;A 640×360 clip at 15 fps for 4 seconds is a few hundred frames — small enough to embed anywhere and still sharp.&lt;/p&gt;

&lt;h2&gt;
  
  
  Step 4 — Convert, then compress in the same pass
&lt;/h2&gt;

&lt;p&gt;The workflow that avoids re-uploading: take the trimmed, cropped source, convert it to GIF at your chosen frame rate and size, and if the result is still heavy, run a compression pass on the output. Because the tools share one credit wallet, the convert and the compress draw from the same balance at per-file rates (from 40 credits, priced by output seconds and shown before the run). Failed jobs cost nothing, so you can tune the settings without spending credits.&lt;/p&gt;

&lt;p&gt;If you need the clip for the web rather than a GIF, export &lt;strong&gt;VP9 / WebM&lt;/strong&gt; instead: a typical MP4-to-WebM pass comes out roughly &lt;strong&gt;4× smaller&lt;/strong&gt; than the GIF at the same visual quality, and it scales to any size. GIF is for ubiquitous embedding; WebM is for actually shipping small.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F5suqcbukhqfyhgw2wehx.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F5suqcbukhqfyhgw2wehx.png" alt=" " width="800" height="450"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F3bjvxs2ffneplup04elw.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F3bjvxs2ffneplup04elw.png" alt=" " width="800" height="450"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  A note on privacy
&lt;/h2&gt;

&lt;p&gt;Clips you upload for conversion are deleted when the job ends, results are kept 24 hours, and nothing is used to train models. For internal demos and support clips that matters more than people expect — your footage does not linger on someone else's server.&lt;/p&gt;

&lt;h2&gt;
  
  
  The takeaway
&lt;/h2&gt;

&lt;p&gt;Good GIFs are made before the conversion, not after. Trim to the moment, crop to the action, set the frame rate to the lowest value that still reads (15 fps is the usual answer), and cap the dimensions. Then convert and compress in one session on a per-file balance that never expires. You get a crisp loop that actually embeds instead of a 14 MB smear.&lt;/p&gt;

&lt;p&gt;Try the workflow on your own clips: &lt;strong&gt;&lt;a href="https://reeloria.com" rel="noopener noreferrer"&gt;reeloria.com&lt;/a&gt;&lt;/strong&gt; — five browser video tools (video to GIF, compress, convert, trim, burn-in subtitles) on one credit balance, priced per file with the cost shown before every run.&lt;/p&gt;

</description>
      <category>saas</category>
      <category>productivity</category>
      <category>webdev</category>
      <category>typescript</category>
    </item>
    <item>
      <title>Background Removal Without a Subscription: A 3-Step Workflow</title>
      <dc:creator>disankyo</dc:creator>
      <pubDate>Wed, 04 Nov 2026 06:52:14 +0000</pubDate>
      <link>https://dev.to/disankyo/background-removal-without-a-subscription-a-3-step-workflow-3fp6</link>
      <guid>https://dev.to/disankyo/background-removal-without-a-subscription-a-3-step-workflow-3fp6</guid>
      <description>&lt;p&gt;Every product photo, profile picture, and marketplace listing seems to come with the same small task: removing the background.&lt;/p&gt;

&lt;p&gt;There are plenty of ways to do it. You can open a desktop editor, spend time tracing around the subject, or sign up for an online tool that charges a monthly subscription. But if you only need to process images occasionally, neither option is especially convenient.&lt;/p&gt;

&lt;p&gt;A simpler workflow is to upload the image, remove the background, and download the result. You pay for the images you process rather than keeping another subscription running in the background.&lt;/p&gt;

&lt;p&gt;Here’s how it works.&lt;/p&gt;

&lt;h2&gt;
  
  
  Why pay per image?
&lt;/h2&gt;

&lt;p&gt;Background removal is often something people do in batches. You might have 40 product photos to prepare this week, then not touch another image for a month. Or you might need a few headshots before an event and nothing else afterward.&lt;/p&gt;

&lt;p&gt;That makes a monthly subscription hard to justify for occasional use.&lt;/p&gt;

&lt;p&gt;With a pay-per-image setup, you can process a batch when you need to and leave the rest of your credits alone. The price is shown before the job starts, and unused credits stay in your account rather than disappearing at the end of the month.&lt;/p&gt;

&lt;p&gt;The goal is simple: get a clean result when you need one, without paying for software while you're not using it.&lt;/p&gt;

&lt;h2&gt;
  
  
  Step 1: Start with a decent source image
&lt;/h2&gt;

&lt;p&gt;Background removal tools can handle a lot, but the quality of the original photo still matters.&lt;/p&gt;

&lt;p&gt;A few things help:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Use a reasonably well-lit photo.&lt;/strong&gt; The lighting doesn't have to be perfect, but a clearly visible subject gives the model more to work with.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Make sure the subject stands out from the background.&lt;/strong&gt; A white cat against a white wall, for example, is difficult for any background removal tool.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Use the highest-resolution version you have.&lt;/strong&gt; It's usually better to start with a sharp image and make adjustments afterward.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Once the image is ready, upload it and run the background removal.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/PLACEHOLDER_before.jpg" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/PLACEHOLDER_before.jpg" alt="Before: a product shot on a busy background" width="800" height="400"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  Step 2: Remove the background
&lt;/h2&gt;

&lt;p&gt;This is the part that used to involve selection tools, masks, and a lot of zooming in around the edges.&lt;/p&gt;

&lt;p&gt;Quixary handles the processing automatically and routes jobs through its available image models. If one provider has a problem during processing, the system can switch providers without requiring you to start the job over.&lt;/p&gt;

&lt;p&gt;Depending on what you need, you can export:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;A &lt;strong&gt;transparent PNG&lt;/strong&gt; for products, logos, stickers, and other images that need to sit on a new background.&lt;/li&gt;
&lt;li&gt;An image with a &lt;strong&gt;plain white background&lt;/strong&gt;, which is useful for many product listings and marketplaces.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Modern image models are also better at some of the awkward cases that used to require manual cleanup:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Hair and fur&lt;/li&gt;
&lt;li&gt;Glass and other semi-transparent objects&lt;/li&gt;
&lt;li&gt;Detailed outlines, such as bicycles, plants, or people in motion&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Results will still vary depending on the original image. If an edge looks wrong or part of the subject is missing, try using a higher-quality source image or adjusting the crop before running it again.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/PLACEHOLDER_after.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/PLACEHOLDER_after.png" alt="After: clean transparent cutout on checkerboard" width="800" height="400"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  Step 3: Keep editing if you need to
&lt;/h2&gt;

&lt;p&gt;Removing the background is often just one part of preparing an image.&lt;/p&gt;

&lt;p&gt;After the cutout, you can continue working with the result in the same session. Depending on what the image needs, you can:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Upscale it&lt;/strong&gt; for larger displays or print.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Adjust the lighting&lt;/strong&gt; if the original image looks flat.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Extend the canvas&lt;/strong&gt; when you need extra space around the subject.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Retouch details&lt;/strong&gt; before exporting the final version.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;The tools use the same credit balance, so you don't need separate subscriptions or accounts for each step. The cost of each operation is shown before you run it.&lt;/p&gt;

&lt;h2&gt;
  
  
  When a subscription might make more sense
&lt;/h2&gt;

&lt;p&gt;Paying per image isn't the cheapest option for everyone.&lt;/p&gt;

&lt;p&gt;If you're processing hundreds or thousands of images every month, a subscription with a high or unlimited usage allowance may work out better. A photo studio or a business with a constant stream of product images might prefer that model.&lt;/p&gt;

&lt;p&gt;For occasional or irregular use, though, paying only when you have images to process can be easier to manage.&lt;/p&gt;

&lt;h2&gt;
  
  
  The bottom line
&lt;/h2&gt;

&lt;p&gt;For most background removal jobs, the workflow is straightforward: upload the image, remove the background, and download the result.&lt;/p&gt;

&lt;p&gt;If the image needs more work, you can continue with upscaling, relighting, or other edits without starting over in another tool.&lt;/p&gt;

&lt;p&gt;If you'd like to try it with your own images, visit &lt;strong&gt;quixary.com&lt;/strong&gt;. The platform brings several image editing tools together under one credit balance, with the cost shown before each image is processed.&lt;/p&gt;

</description>
      <category>ai</category>
      <category>voice</category>
      <category>productivity</category>
      <category>tooling</category>
    </item>
    <item>
      <title>《What I Learned Building a Pay-Per-Use Voice Cloning Platform》</title>
      <dc:creator>disankyo</dc:creator>
      <pubDate>Tue, 03 Nov 2026 04:06:58 +0000</pubDate>
      <link>https://dev.to/disankyo/-3bm6</link>
      <guid>https://dev.to/disankyo/-3bm6</guid>
      <description>&lt;p&gt;What I Learned Building a Pay-Per-Use Voice Cloning Platform&lt;/p&gt;

&lt;p&gt;Every AI voice tool I tried had the same shape: pick a plan, $10–30/month, cancel anytime (you won't). Before you even know if the voice quality works for your project, you've committed to a month. I kept thinking: this pricing model fits SaaS dashboards, not creative tools you use in bursts.&lt;/p&gt;

&lt;p&gt;So I built Resona the other way around — credits instead of a plan.&lt;/p&gt;

&lt;p&gt;The model, in one paragraph&lt;/p&gt;

&lt;p&gt;You buy credits ($1 = 50 credits) and they never expire. Cloning a voice costs 100 credits. Generating speech costs 1,500 credits per 1M UTF-8 bytes of text — which works out to fractions of a cent for a typical paragraph. If you clone one voice and narrate one video this month, you pay for exactly that. If you don't touch it for two months, nothing is wasted.&lt;/p&gt;

&lt;p&gt;Three things I learned shipping it&lt;/p&gt;

&lt;ol&gt;
&lt;li&gt;&lt;p&gt;Sample quality beats sample quantity. People email us asking why their clone sounds off, and it's almost never the algorithm — it's a phone recording with room echo. Thirty seconds of clean audio from one consistent mic beats five messy minutes. Record in a soft-furnished room, keep the mic distance fixed, and don't mix takes from different days.&lt;/p&gt;&lt;/li&gt;
&lt;li&gt;&lt;p&gt;Billing by bytes changes how people write. When you pay per minute of studio time or per month of access, nobody thinks about the text. When you pay per byte, people notice that "um, so, yeah, basically..." costs money. The side effect is better scripts — tighter sentences, fewer filler words. I didn't expect a pricing model to act as an editor, but here we are.&lt;/p&gt;&lt;/li&gt;
&lt;li&gt;&lt;p&gt;Multilingual is not a checkbox. Resona generates speech in 83 languages, and the requests we didn't anticipate are the fun ones: the same cloned voice reading a script in English and then in Japanese for a localized tutorial. Keeping a voice "recognizably the same person" across languages turned out to be the feature people care about most — more than raw audio fidelity.&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;What's next&lt;/p&gt;

&lt;p&gt;Voice design (describing a voice in text and generating it), transcription with speaker separation, and a public library of 200,000+ ready-to-use voices. The roadmap is driven by what burst-users ask for, because they're the ones the pricing model was built for.&lt;/p&gt;

&lt;p&gt;Full disclosure: I'm the founder of Resona (resonaio.com), so obviously biased toward pay-per-use — but the per-byte vs per-month math above holds no matter which vendor you pick. If you try it, I'd genuinely like to hear where the model breaks for your use case.&lt;/p&gt;

</description>
      <category>ai</category>
      <category>voice</category>
      <category>productivity</category>
      <category>saas</category>
    </item>
    <item>
      <title>Word to PDF Without Losing Formatting — What Actually Works</title>
      <dc:creator>disankyo</dc:creator>
      <pubDate>Mon, 19 Oct 2026 03:51:50 +0000</pubDate>
      <link>https://dev.to/disankyo/word-to-pdf-without-losing-formatting-what-actually-works-4hkb</link>
      <guid>https://dev.to/disankyo/word-to-pdf-without-losing-formatting-what-actually-works-4hkb</guid>
      <description>&lt;p&gt;Word to PDF Without Losing Formatting — What Actually Works&lt;/p&gt;

&lt;p&gt;Nothing's more annoying than a converted PDF where the margins drift and&lt;br&gt;
the fonts swap. Here's what preserves fidelity:&lt;/p&gt;

&lt;ol&gt;
&lt;li&gt;LibreOffice "Export as PDF" — gold standard offline, respects layout.&lt;/li&gt;
&lt;li&gt;Microsoft Word's own "Save as PDF" — fine for simple docs.&lt;/li&gt;
&lt;li&gt;A web converter that uses a real layout engine, e.g.
&lt;a href="https://www.master-converter.com/en/tools/word-pdf?utm_source=devto&amp;amp;utm_medium=article&amp;amp;utm_campaign=social_seo_w2" rel="noopener noreferrer"&gt;https://www.master-converter.com/en/tools/word-pdf?utm_source=devto&amp;amp;utm_medium=article&amp;amp;utm_campaign=social_seo_w2&lt;/a&gt;
(pay-per-file, no sub — good if you only convert occasionally)&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;Common gotchas:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Embedded fonts missing on the source machine → substitution. Always embed fonts.&lt;/li&gt;
&lt;li&gt;Complex tables → test a sample page first.&lt;/li&gt;
&lt;li&gt;Very large docs → split then convert if a tool times out.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;If formatting breaks, fall back to Option 1. It rarely does.&lt;/p&gt;

</description>
      <category>productivity</category>
      <category>pdf</category>
      <category>saas</category>
      <category>tools</category>
    </item>
  </channel>
</rss>
