DEV Community: WEI LIU

Qwen Image 2.1: More Control for AI Image Creation and Editing

WEI LIU — Fri, 02 Oct 2026 01:18:38 +0000

Developers working with image generation often need more control than a single prompt can provide.

Qwen Image 2.1 is an AI model for text-to-image creation and image editing. Its documented capabilities include detailed prompts, multiple reference images, portrait and typography refinement, and native transparent RGBA images.

That combination can be useful when prototyping visual interfaces, iterating on product concepts, or preparing assets that need transparent backgrounds.

A practical workflow

Start with a detailed prompt that describes the visual goal.
Add reference images when preserving a visual direction matters.
Refine portraits or typography in a focused iteration instead of rebuilding the whole concept.
Export transparent RGBA images when the asset needs to move cleanly between tools or layers.

The important idea is controllability: image generation becomes more useful when the workflow supports precise revisions, not only first-pass exploration.

Learn more about Qwen Image 2.1.

From Written Ideas to Cinematic Motion with Opus55video

WEI LIU — Thu, 01 Oct 2026 06:17:34 +0000

A written idea can already contain the beginning of a shot: a subject, a setting, an action, and a mood. Opus55video turns that starting point into a short video workflow.

Start with the scene

Opus55video is an online workspace for creating videos from text. Describe the scene you want to see, then add the details that matter to the direction: who or what is in the frame, where the action happens, and how the moment should feel.

Direct the format and timing

Once the idea is clear, choose the format and duration that fit the destination. Landscape, portrait, and square framing can serve different contexts, from a wide concept presentation to a mobile-first story. A focused duration also helps keep the shot purposeful.

Refine the direction

The creative direction stays with you. Try a different setting, camera movement, lighting choice, or pace, then compare the result with the original idea. The goal is a simple path from written prompt to a visual reference that can be watched, discussed, and refined.

Opus55video is built for creative exploration, storyboard references, social content ideas, and conversations about visual direction. Start with one clear scene and let the next frame follow from the words.

hotel-lobby: A Warm, Two-Person AI Video Concept

WEI LIU — Thu, 01 Oct 2026 05:29:45 +0000

hotel-lobby is a freemium AI video concept set in a warm orange recording booth. It features two performers sharing one microphone in a natural call-and-response scene.

Its tagline is “Two Faces. One Mic. Endless Chemistry.” Learn more at hotel-lobby.

I Built an AI Video Tool. The API Was the Easy Part

WEI LIU — Fri, 18 Sep 2026 12:28:18 +0000

A few weeks ago, I started building a small AI video project.

My assumption was pretty simple: if video generation becomes fast enough, there should be interesting products you can build around it.

So I connected a video generation API, built a basic interface, and got the first version working.

Prompt in. Video out.

At first, that felt like progress.

Then I actually tried using it as a real workflow.

And I realized the API was probably the easiest part of the whole product.

Generating one video is easy

Most AI video demos follow the same pattern:

Prompt → Generate → Wait → Video

That's fine when you're testing a model.

But imagine you want to create short-form content regularly.

Now the workflow becomes something like:

Idea → Prompt → Generate → Wait → Review → Download → Organize → Upload → Publish

Then you do it again.

And again.

If you want to create 10 videos, you don't really have a video generation problem anymore.

You have a workflow problem.

That distinction wasn't obvious to me when I started building.

Waiting becomes a UX problem

Another thing I underestimated was generation time.

With many normal web applications, a user clicks something and expects an almost immediate response.

AI video doesn't always work like that.

A generation can take long enough that keeping someone on a loading screen feels wrong.

My first version basically did this:

click → wait → result

It worked technically, but it didn't feel like a good product.

I started thinking about each generation as a job instead.

submit → queue → process → result

That small change affects a lot of things.

Now multiple generations can be submitted without forcing the user to wait for each one individually.

It also makes batch generation much easier to think about.

The next problem appears after the video is generated

Once I had generation working, another question appeared:

What happens after the video exists?

If the user still needs to download every video, rename files, open another platform, upload them manually, and repeat the process, I've only automated one part of the job.

This is where my original idea started changing.

Instead of building another page with a prompt box and a Generate button, I became more interested in the entire pipeline:

Create → Queue → Review → Publish

I'm currently testing this idea through a small project called ReelsLaunch.

I'm using it to experiment with short AI video generation, queues, batch workflows, and reducing the manual steps between creating a video and publishing it.

It's still early, and I'm not sure yet which part of that workflow will turn out to be the most important.

But building it has already changed how I think about AI products.

Three things I've learned so far

Don't build around the model

It's tempting to see a new model and immediately think:

I should build a website for this model.

I did something close to that at the beginning.

But models change very quickly.

Today's impressive model can become a standard feature somewhere else a few months later.

The model is infrastructure.

The more interesting question is what workflow becomes possible because that infrastructure now exists.

Long-running AI tasks need different UX

A traditional request/response interface isn't always a good fit for video generation.

Users shouldn't have to babysit a generation.

For longer AI tasks, queues, background processing, status tracking, retries, and notifications can become just as important as the generation itself.

This sounds obvious in hindsight.

It wasn't obvious to me when I started.

Look at what happens before and after the AI

This is probably the biggest lesson for me.

When we build AI products, it's easy to focus almost entirely on the AI step.

But users usually have a larger job they are trying to complete.

For short-form video, generation is only one piece.

There is planning before it.

There is reviewing, organizing, and publishing after it.

The opportunity may not be in making the Generate button better.

It may be in removing the need to click five other buttons afterward.

I'm still figuring this out

I don't know yet whether the best version of this product is primarily a video generator, a publishing tool, or something closer to an automated content pipeline.

That's part of why I'm building it.

But one question has become much more useful to me than asking which AI model I should use:

What repetitive work can disappear because this model exists?

If you're building something on top of an AI API, I'd be interested to hear how you think about this.

Are you building around the model itself, or around the workflow that happens before and after it?

The True Cost of Cinematic AI Video: Navigating MiniMax H3 Max API Pricing

WEI LIU — Mon, 07 Sep 2026 03:08:14 +0000

ntroduction
The AI video landscape is shifting rapidly. With the recent release of MiniMax H3 Max on platforms like fal.ai, we are finally seeing "cinematic" quality that rivals Hollywood storyboards. However, as these models move from research labs to production environments, a new challenge emerges: Cost Transparency.
Unlike flat-rate subscriptions, high-end models use a "per-second" billing model that can lead to significant "sticker shock" if not planned correctly. I’ve been tracking these shifts and built h3price.com to help creators navigate this new economy.
Why the "Per-Second" Model Changes Everything
Most creators are used to monthly credits, but the H3 Max API is billed based on output duration and resolution. Here’s the current breakdown for regular rates:

480p Resolution: $0.05 per second.
768p Resolution: $0.08 per second.

At first glance, $0.08 sounds small. But for a professional 60-second sequence at 768p, you are looking at $4.80 per minute. When you factor in the "trial and error" inherent in AI generation, costs can escalate quickly.
Optimization is the New Skill
To help the community manage these workflows, I developed a few tools on h3price.com:

H3 Cost Calculator: Essential for budgeting production runs. You can estimate total spend by resolution and clip volume before hitting "Generate."
Cinematic Prompt Generator: Since every second costs money, your first prompt needs to be your best. I’ve built a generator specifically tuned for H3 Max’s native understanding of camera movement and lighting.
Promo Tracking: We monitor launch offers (like the current 50% discount for the first 14 days) to ensure you're timing your heavy production runs during discount windows.

The Verdict
MiniMax H3 Max is a game-changer for visual storytelling, but it requires a "Developer Mindset" toward budgeting. Whether you are a solo creator or a creative agency, understanding the granular cost of your frames is the only way to scale.
Check out the tools and pricing guide here: https://www.h3price.com
I’d love to hear from other devs—how are you handling the shift to per-second API billing in your video workflows?

Stop Fighting the Filters: A Free, Browser-Based Uncensored AI Image Editor

WEI LIU — Wed, 02 Sep 2026 05:30:17 +0000

As developers and creators, we’ve all been there: you have a specific creative vision, but the mainstream AI image generators block your prompt with a "content filter" or force you to sign up for a subscription just to see a single draft.
I recently came across a tool that takes a different approach: Uncensored AI Editor. It’s designed to be a direct, friction-free workspace for turning text into high-quality visuals without the usual gatekeeping.
What makes it different?
Unlike Midjourney or DALL-E, this editor focuses on momentum and privacy:

No Sign-Up Required: You can open the browser and start generating immediately. No email verification, no social logins.
Truly Free & No Watermarks: The images you download are clean. There are no semi-transparent logos covering your work, and you don't need a credit card to get started.
Browser-Based Workflow: It runs entirely in your browser (phone, tablet, or laptop), making it a great tool for quick mockups or social media assets on the go.
Reference Image Support: You can upload your own images to guide the color palette, composition, or style, which is a game-changer for maintaining brand consistency.

Technical Highlights
For those interested in the workflow, the editor provides significant control over the output:

Resolution Control: Choose between 1K and 2K outputs depending on whether you need a quick draft or a production-ready asset.
Aspect Ratio Variety: It supports 13 native aspect ratios, covering everything from mobile stories to cinematic presentations.
Transparent Credits: Every task shows the credit cost upfront, so you know exactly what you’re spending before hitting generate.

Why "Uncensored"?
The "uncensored" tag here means the tool avoids the blanket creative filters that often stifle legitimate artistic exploration. While it still enforces safety standards (no illegal content), it doesn't interrupt your ordinary creative process with overly sensitive keyword blocks.
If you're looking for a new tool to add to your design stack—especially for landing page hero images, product mockups, or concept visualization—give it a try.
Check it out here: https://www.uncensoredaieditor.com