<?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: liangqi zhao</title>
    <description>The latest articles on DEV Community by liangqi zhao (@liangqi_zhao_b047bd1f3236).</description>
    <link>https://dev.to/liangqi_zhao_b047bd1f3236</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%2F4211870%2F5ac422a1-b72d-4d65-a9a3-ca511ffa18c0.png</url>
      <title>DEV Community: liangqi zhao</title>
      <link>https://dev.to/liangqi_zhao_b047bd1f3236</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/liangqi_zhao_b047bd1f3236"/>
    <language>en</language>
    <item>
      <title>A small set of GIF editing tools for browser workflows</title>
      <dc:creator>liangqi zhao</dc:creator>
      <pubDate>Tue, 17 Nov 2026 11:21:57 +0000</pubDate>
      <link>https://dev.to/liangqi_zhao_b047bd1f3236/a-small-set-of-gif-editing-tools-for-browser-workflows-5007</link>
      <guid>https://dev.to/liangqi_zhao_b047bd1f3236/a-small-set-of-gif-editing-tools-for-browser-workflows-5007</guid>
      <description>&lt;p&gt;When I work with animated GIFs, the same small adjustments come up repeatedly: reduce file size, change playback speed, or remove and reorder frames. I put those tasks together as browser tools on Animated GIF Generator.&lt;/p&gt;

&lt;p&gt;The compressor reduces GIF file size. The speed changer adjusts playback speed while preserving the image data. The frame editor can remove, reorder, and retime frames. The tools run in the browser without uploading the file.&lt;/p&gt;

&lt;p&gt;AI GIF generation is currently offline, so the available tools are for editing existing GIFs.&lt;/p&gt;

&lt;p&gt;Tools: &lt;a href="https://animatedgifgenerator.com/" rel="noopener noreferrer"&gt;Animated GIF Generator&lt;/a&gt;&lt;/p&gt;

</description>
    </item>
    <item>
      <title>Removing a watermark without cropping the frame</title>
      <dc:creator>liangqi zhao</dc:creator>
      <pubDate>Tue, 17 Nov 2026 10:29:46 +0000</pubDate>
      <link>https://dev.to/liangqi_zhao_b047bd1f3236/removing-a-watermark-without-cropping-the-frame-3a7j</link>
      <guid>https://dev.to/liangqi_zhao_b047bd1f3236/removing-a-watermark-without-cropping-the-frame-3a7j</guid>
      <description>&lt;p&gt;When a logo or burned-in text sits over footage, cropping can hide it, but it also changes the composition and may cut out useful detail. Inpainting is another option: mark the region, then reconstruct the pixels behind it from surrounding frames.&lt;/p&gt;

&lt;p&gt;I built WatermarkRemoverVideo for cleaning up videos you own or have permission to edit. It accepts MP4 files up to 40 MB and 10 minutes. You select the watermark, logo, or text region; the tool estimates what was behind it and fills the area while keeping the original frame dimensions.&lt;/p&gt;

&lt;p&gt;A few practical checks help set expectations:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Review the whole clip, especially where the background moves behind the selected area.&lt;/li&gt;
&lt;li&gt;Use a tight selection around the mark. A larger selection can remove more context than intended.&lt;/li&gt;
&lt;li&gt;Compare several frames before exporting. Reconstructed pixels are an estimate, not recovery of the original source data.&lt;/li&gt;
&lt;li&gt;If the mark covers a moving subject or a highly detailed pattern, results may be less consistent. Cropping or finding an unmarked source may be a better fit.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;For clips that belong to you, the tool is at &lt;a href="https://watermarkremovervideo.com/" rel="noopener noreferrer"&gt;https://watermarkremovervideo.com/&lt;/a&gt;. It charges 10 credits per second of video. I made it to preserve the full frame when a straightforward crop would lose part of the shot.&lt;/p&gt;

</description>
      <category>webdev</category>
    </item>
    <item>
      <title>Turning handwriting into a TTF font with Inkown</title>
      <dc:creator>liangqi zhao</dc:creator>
      <pubDate>Tue, 17 Nov 2026 08:46:57 +0000</pubDate>
      <link>https://dev.to/liangqi_zhao_b047bd1f3236/turning-handwriting-into-a-ttf-font-with-inkown-36en</link>
      <guid>https://dev.to/liangqi_zhao_b047bd1f3236/turning-handwriting-into-a-ttf-font-with-inkown-36en</guid>
      <description>&lt;p&gt;I built Inkown to turn a handwriting sample into an installable TTF font. The flow accepts JPG, PNG, and WebP images, then generates a preview before download. New accounts get two AI generations for preview.&lt;/p&gt;

&lt;p&gt;The current output covers 76 basic Latin characters plus a space. Accented characters and cursive joins are not supported, so a clean sample with separated letters works best. The downloadable font is a one-time purchase.&lt;/p&gt;

&lt;p&gt;Product: &lt;a href="https://inkown.com/" rel="noopener noreferrer"&gt;Inkown&lt;/a&gt;&lt;/p&gt;

</description>
      <category>webdev</category>
    </item>
    <item>
      <title>Designing a Narrow Photo-to-Video Workflow for a Meme-Inspired Tool</title>
      <dc:creator>liangqi zhao</dc:creator>
      <pubDate>Mon, 16 Nov 2026 10:17:54 +0000</pubDate>
      <link>https://dev.to/liangqi_zhao_b047bd1f3236/designing-a-narrow-photo-to-video-workflow-for-a-meme-inspired-tool-290b</link>
      <guid>https://dev.to/liangqi_zhao_b047bd1f3236/designing-a-narrow-photo-to-video-workflow-for-a-meme-inspired-tool-290b</guid>
      <description>&lt;h2&gt;
  
  
  Start with a clear input contract
&lt;/h2&gt;

&lt;p&gt;I built &lt;a href="https://talktomedoja.com/" rel="noopener noreferrer"&gt;Talk to Me Doja AI&lt;/a&gt; around one simple input: a clear photo of one adult that the uploader has permission to use. The tool makes an AI interpretation of the shy, hands-under-the-chin sway associated with the Talk to Me Doja meme, so there is no need to film the dance.&lt;/p&gt;

&lt;p&gt;The upload accepts JPG, PNG, or WebP images up to 10 MB. A recent, evenly lit, front-facing head-and-shoulders photo is a better starting point; group photos, sunglasses, and heavy filters can reduce quality.&lt;/p&gt;

&lt;h2&gt;
  
  
  Describe the output as a request, not a guarantee
&lt;/h2&gt;

&lt;p&gt;A generation request is for a 14-second vertical clip. A successful result is an MP4, but actual duration, framing, likeness, and movement can vary. The 9:16 reference is a starting point, not a fixed output promise. I recommend reviewing the file before sharing it.&lt;/p&gt;

&lt;p&gt;This distinction matters for any photo-to-video product: an AI result is an interpretation, not a recording of the person performing the move.&lt;/p&gt;

&lt;h2&gt;
  
  
  Make the cost visible before generation
&lt;/h2&gt;

&lt;p&gt;One 14-second request costs 100 credits. Packs are $9.99 for 100 credits, $24.99 for 300, and $39.99 for 500. There is no subscription or free generation allowance. Pack credits do not expire while the service operates, and credits are returned if a video cannot be created.&lt;/p&gt;

&lt;p&gt;Showing the cost before the user starts a job makes the trade-off clear and avoids treating a paid generation as a surprise.&lt;/p&gt;

&lt;h2&gt;
  
  
  Keep permission and affiliation clear
&lt;/h2&gt;

&lt;p&gt;The uploader must be an adult and have permission to use the photo. The product does not accept photos of minors, celebrities, or public figures. Talk to Me Doja AI is an independent tool; it is not an official Doja Cat or anime release. Do not assume a generated clip includes licensed music—check the destination platform’s rules before posting.&lt;/p&gt;

&lt;p&gt;That is the scope: one permitted photo, a clearly priced request, and an MP4 to review if generation succeeds.&lt;/p&gt;

</description>
      <category>ai</category>
      <category>webdev</category>
    </item>
    <item>
      <title>Building a photo-to-video flow with clear expectations</title>
      <dc:creator>liangqi zhao</dc:creator>
      <pubDate>Mon, 16 Nov 2026 04:34:05 +0000</pubDate>
      <link>https://dev.to/liangqi_zhao_b047bd1f3236/building-a-photo-to-video-flow-with-clear-expectations-4ai7</link>
      <guid>https://dev.to/liangqi_zhao_b047bd1f3236/building-a-photo-to-video-flow-with-clear-expectations-4ai7</guid>
      <description>&lt;p&gt;I built &lt;a href="https://talktomedoja.com/" rel="noopener noreferrer"&gt;Talk to Me Doja AI&lt;/a&gt; as an independent photo-to-video tool inspired by the “Talk to Me Doja” meme. The goal is straightforward: start with one permitted portrait instead of filming the dance, then review the generated clip before sharing it.&lt;/p&gt;

&lt;p&gt;The input rules matter more than clever prompts. The tool accepts one clear, front-facing adult photo in JPG, PNG, or WebP, up to 10 MB. Even lighting and a head-and-shoulders crop are a better starting point; group photos, sunglasses, and heavy filters can reduce quality.&lt;/p&gt;

&lt;p&gt;A 14-second request costs 100 credits. Successful generations are MP4 files, but actual duration, framing, and visual similarity can vary. The tool is not affiliated with Doja Cat or the anime referenced by the meme, and it does not include licensed music by default. I treat those limits as part of the product workflow: set expectations before generation and review the output before posting.&lt;/p&gt;

&lt;p&gt;Credits are sold in one-time packs rather than a subscription. If a generation fails, the credits are refunded. Uploaded photos may be automatically deleted seven days after upload once they are no longer needed for an in-progress video; completed videos are kept for 30 days.&lt;/p&gt;

&lt;p&gt;If you want to try the workflow, the product details and photo requirements are at &lt;a href="https://talktomedoja.com/" rel="noopener noreferrer"&gt;talktomedoja.com&lt;/a&gt;.&lt;/p&gt;

</description>
      <category>webdev</category>
    </item>
    <item>
      <title>A small QA checklist for photo-to-video generators</title>
      <dc:creator>liangqi zhao</dc:creator>
      <pubDate>Mon, 16 Nov 2026 01:54:04 +0000</pubDate>
      <link>https://dev.to/liangqi_zhao_b047bd1f3236/a-small-qa-checklist-for-photo-to-video-generators-5a3b</link>
      <guid>https://dev.to/liangqi_zhao_b047bd1f3236/a-small-qa-checklist-for-photo-to-video-generators-5a3b</guid>
      <description>&lt;p&gt;A photo-to-video flow is hard to evaluate if each test changes the input, crop, and request at once. A small repeatable checklist helps separate input problems from output variation.&lt;/p&gt;

&lt;h2&gt;
  
  
  1. Freeze the test input
&lt;/h2&gt;

&lt;p&gt;Use a portrait you have permission to process. Record the file format, dimensions, and whether the face is centered and unobstructed. Keep the same source image for comparisons. If the input changes between runs, the output comparison is weaker.&lt;/p&gt;

&lt;h2&gt;
  
  
  2. Check the framing contract
&lt;/h2&gt;

&lt;p&gt;Write down the requested output shape and duration before you generate. For a vertical clip, inspect the full frame at phone size: is the face still visible, is there enough headroom, and does the motion leave the safe area? A preview can look fine on desktop while cropping poorly on a narrow screen.&lt;/p&gt;

&lt;h2&gt;
  
  
  3. Review the artifact, not just the preview
&lt;/h2&gt;

&lt;p&gt;Open the exported file. Confirm it plays, has the expected orientation and duration, and contains the motion you asked for. Look for facial distortion, sudden jumps, or a crop that hides the subject. Treat each result as a sample: generation can vary even when the input is unchanged.&lt;/p&gt;

&lt;h2&gt;
  
  
  4. Keep the test honest
&lt;/h2&gt;

&lt;p&gt;Record the input and request settings, then note what passed and failed. Avoid calling one successful sample a guarantee. Before sharing a clip, check rights and consent for the source photo and make sure the result fits its intended audience.&lt;/p&gt;

&lt;p&gt;As one concrete example, Talk to Me Doja AI accepts one portrait and requests a 14-second vertical video. A generation costs 100 credits; successful outputs are MP4 files, and results vary with the photo and generation. It is an independent product, not affiliated with Doja Cat or any rights holder.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://talktomedoja.com/" rel="noopener noreferrer"&gt;Try Talk to Me Doja AI&lt;/a&gt;&lt;/p&gt;

</description>
      <category>video</category>
    </item>
    <item>
      <title>A Fixed Baseline Makes Review Faster</title>
      <dc:creator>liangqi zhao</dc:creator>
      <pubDate>Fri, 13 Nov 2026 05:49:48 +0000</pubDate>
      <link>https://dev.to/liangqi_zhao_b047bd1f3236/a-fixed-baseline-makes-review-faster-48b8</link>
      <guid>https://dev.to/liangqi_zhao_b047bd1f3236/a-fixed-baseline-makes-review-faster-48b8</guid>
      <description>&lt;p&gt;Review is fastest when the comparison target is fixed. Write the brief and the reference details first — product identity, allowed variations, must-stay-fixed details — then judge every generated result against that written baseline instead of against your memory of the last image you liked.&lt;/p&gt;

&lt;p&gt;Published by &lt;a href="https://jynno.com" rel="noopener noreferrer"&gt;Jynno&lt;/a&gt;, an AI creative studio for image generation, photo editing and video creation. More guides: &lt;a href="https://jynno.com/" rel="noopener noreferrer"&gt;https://jynno.com/&lt;/a&gt;&lt;/p&gt;

</description>
      <category>aiworkflowbeginners</category>
    </item>
    <item>
      <title>Write the Reference Sheet Before the First Prompt</title>
      <dc:creator>liangqi zhao</dc:creator>
      <pubDate>Fri, 13 Nov 2026 05:34:30 +0000</pubDate>
      <link>https://dev.to/liangqi_zhao_b047bd1f3236/write-the-reference-sheet-before-the-first-prompt-3528</link>
      <guid>https://dev.to/liangqi_zhao_b047bd1f3236/write-the-reference-sheet-before-the-first-prompt-3528</guid>
      <description>&lt;p&gt;Before generating character scenes, write a short reference sheet: face shape, hair, outfit colors, signature accessories. The sheet is cheaper to check against than a rendered image, and it gives every later comparison a fixed baseline. Compare each new frame to the sheet, not to the previous frame, so drift does not accumulate.&lt;/p&gt;

&lt;p&gt;Published by &lt;a href="https://jynno.com" rel="noopener noreferrer"&gt;Jynno&lt;/a&gt;, an AI creative studio for image generation, photo editing and video creation. More guides: &lt;a href="https://jynno.com/" rel="noopener noreferrer"&gt;https://jynno.com/&lt;/a&gt;&lt;/p&gt;

</description>
      <category>aiwritingtutorial</category>
    </item>
    <item>
      <title>Change One Variable at a Time When Iterating on Product Images</title>
      <dc:creator>liangqi zhao</dc:creator>
      <pubDate>Fri, 13 Nov 2026 05:14:14 +0000</pubDate>
      <link>https://dev.to/liangqi_zhao_b047bd1f3236/change-one-variable-at-a-time-when-iterating-on-product-images-2lab</link>
      <guid>https://dev.to/liangqi_zhao_b047bd1f3236/change-one-variable-at-a-time-when-iterating-on-product-images-2lab</guid>
      <description>&lt;p&gt;When iterating on generated product images, change one major variable at a time — background, lighting direction or camera distance — so you can tell which choice actually improved the result. Keep product silhouette, color, material and logo fixed across every variation, and review each result against the original photograph.&lt;/p&gt;

&lt;p&gt;Published by &lt;a href="https://jynno.com" rel="noopener noreferrer"&gt;Jynno&lt;/a&gt;, an AI creative studio for image generation, photo editing and video creation. More guides: &lt;a href="https://jynno.com/" rel="noopener noreferrer"&gt;https://jynno.com/&lt;/a&gt;&lt;/p&gt;

</description>
      <category>aiproductivitymarketing</category>
    </item>
    <item>
      <title>Why Context Scenes Beat Busy Backgrounds for Product Visuals</title>
      <dc:creator>liangqi zhao</dc:creator>
      <pubDate>Fri, 13 Nov 2026 05:12:45 +0000</pubDate>
      <link>https://dev.to/liangqi_zhao_b047bd1f3236/why-context-scenes-beat-busy-backgrounds-for-product-visuals-16ga</link>
      <guid>https://dev.to/liangqi_zhao_b047bd1f3236/why-context-scenes-beat-busy-backgrounds-for-product-visuals-16ga</guid>
      <description>&lt;p&gt;A simple product view explains the object; a context scene shows how it fits into everyday life; a short visual story gives it a role in a moment. Decide what may change (background, lighting, camera distance) and what must stay fixed before generating, and change one major variable at a time.&lt;/p&gt;

&lt;p&gt;Published by &lt;a href="https://jynno.com" rel="noopener noreferrer"&gt;Jynno&lt;/a&gt;, an AI creative studio for image generation, photo editing and video creation. More guides: &lt;a href="https://jynno.com/" rel="noopener noreferrer"&gt;https://jynno.com/&lt;/a&gt;&lt;/p&gt;

</description>
      <category>aidesigntutorials</category>
    </item>
    <item>
      <title>Write Down the Event Before You Describe the Camera</title>
      <dc:creator>liangqi zhao</dc:creator>
      <pubDate>Fri, 13 Nov 2026 04:54:33 +0000</pubDate>
      <link>https://dev.to/liangqi_zhao_b047bd1f3236/write-down-the-event-before-you-describe-the-camera-115j</link>
      <guid>https://dev.to/liangqi_zhao_b047bd1f3236/write-down-the-event-before-you-describe-the-camera-115j</guid>
      <description>&lt;p&gt;For short product videos, a single clear event is easier to judge than a long list of unrelated effects. Write down the action before describing the camera, check whether the product remains recognizable throughout the clip, and review the opening frame against the product identity before generating variations.&lt;/p&gt;

&lt;p&gt;Published by &lt;a href="https://jynno.com" rel="noopener noreferrer"&gt;Jynno&lt;/a&gt;, an AI creative studio for image generation, photo editing and video creation. More guides: &lt;a href="https://jynno.com/" rel="noopener noreferrer"&gt;https://jynno.com/&lt;/a&gt;&lt;/p&gt;

</description>
      <category>videoaimarketing</category>
    </item>
    <item>
      <title>Keep the Product Recognizable When You Generate Variations</title>
      <dc:creator>liangqi zhao</dc:creator>
      <pubDate>Fri, 13 Nov 2026 04:18:59 +0000</pubDate>
      <link>https://dev.to/liangqi_zhao_b047bd1f3236/keep-the-product-recognizable-when-you-generate-variations-4dp2</link>
      <guid>https://dev.to/liangqi_zhao_b047bd1f3236/keep-the-product-recognizable-when-you-generate-variations-4dp2</guid>
      <description>&lt;p&gt;Generated variations drift: a handle changes shape, a logo gets redrawn, a color shifts. Note which product details must stay recognizable before generating, keep the original photograph beside each result while reviewing, and reject any variation that changes the product into something the customer would not receive.&lt;/p&gt;

&lt;p&gt;Published by &lt;a href="https://jynno.com" rel="noopener noreferrer"&gt;Jynno&lt;/a&gt;, an AI creative studio for image generation, photo editing and video creation. More planning guides: &lt;a href="https://jynno.com/" rel="noopener noreferrer"&gt;https://jynno.com/&lt;/a&gt;&lt;/p&gt;

</description>
      <category>aimarketingproductivity</category>
    </item>
  </channel>
</rss>
