Generative AI is now part of most digital content workflows. It can speed up production dramatically, but it doesn’t solve every problem. AI outputs often need extra polish before they’re ready for production: a background swapped, a crop refined, or a blurry detail sharpened. Without those fixes, brands risk shipping assets that feel inconsistent or off-brand.
That gap between fast AI output and production-ready quality creates new challenges for developers and creative teams. Developers end up writing brittle scripts to automate edits. Designers and content teams spend hours making small corrections across hundreds of images. The result is slower launches, more back-and-forth, and a heavier load on teams that are already stretched.
Today, Cloudinary is announcing the launch of a set of new image features designed to solve exactly these problems. With new tools for workflow automation, content provenance, and collaborative editing, brands can now scale AI-driven content with greater confidence, authenticity, and efficiency.
Read the full press release here.
One of the more common issues with generative AI is that outputs are occasionally in need of refinements to make them production-ready. For example, a generated product photo might have a distracting background, an awkward crop, or mismatched colors. The brand-damaging potential of one bad output cannot be overstated, even if 99 out 100 outputs were good. As a result, developers haven’t been able to automate GenAI edits at scale because content teams had no practical way to review or refine the results without jumping back into tools like Photoshop.
Cloudinary’s new Refiners tackle those problems directly. They’re prebuilt functions for common fixes:
- Background removal and replacement for cleaner product photos.
- Outpainting to expand an image naturally for new layouts.
- Focus-area cropping to guide how assets are framed.
The advantage is scale. Cloudinary’s new Refiners let creatives quickly polish AI outputs, giving developers the confidence to programmatically automate edits across hundreds or thousands of assets. The result: Every image is consistent, on-brand, and production-ready.
This combination of automation and human review reflects what many analysts now call “human-in-the-loop AI.” It’s not about letting the model make every decision. It’s about automating repeatable tasks while keeping final control in the hands of people who understand the brand. That balance is what makes GenAI practical in production.
Speed and efficiency solve one set of problems, but AI introduces another: authenticity. As manipulated content becomes more common, brands are under pressure to prove their media is genuine and traceable. Regulators are also paying closer attention. In the EU, for example, the proposed AI Act highlights transparency as a key requirement for generative outputs.
Cloudinary now includes built-in support for C2PA (Coalition for Content Provenance and Authenticity), the open standard for certifying digital content provenance. With C2PA, every image can carry tamper-proof content credentials that documents its history: which edits were made, when, and by what tool. That record persists through transformations, creating a tamper-proof chain of custody.
Enabling it is simple. Add the c2pa flag to your transformation URL:
<code>https://res.cloudinary.com/<cloud_name>/image/upload/c2pa/v1/sample.jpg</code>Code language: HTML, XML (xml)Or set it in the Node.js SDK:
<code>cloudinary.image("sample.jpg", { c2pa: true });</code>Code language: HTML, XML (xml)By baking C2PA into image delivery, Cloudinary gives developers a straightforward way to deliver that transparency.
Refinement and provenance are important, but day-to-day workflows also need to be faster and more collaborative. Cloudinary’s new Editor brings developers and creatives into the same environment. Developers can chain together transformations and see live code snippets update as they experiment. Creatives can preview different outputs side by side, compare results, and make adjustments before campaigns launch.
This kind of shared workspace reduces the “in sequence” problem — when developers have to wait for creatives to finish editing, or vice versa. Instead, teams can prototype and iterate together.
The new Auto Enhance feature helps, too. It analyzes an image and applies the best combination of adjustments, from sharpening and upscaling to restoration. It saves time by removing the need for trial-and-error editing. For many teams, Auto Enhance provides a solid baseline image that can then be fine-tuned if needed.
For e-commerce, where high-quality visuals directly impact performance, these efficiencies matter. A Shopify study found that better product images can boost conversion rates by as much as 30%. Faster, more consistent editing means retailers can refresh catalogs more often and deliver better experiences without bottlenecking launches.
Cloudinary has also refined its underlying GenAI model with domain-specific data. That matters because general-purpose AI models often produce inconsistent results in specialized use cases.
In e-commerce, for instance, product photos usually require clean, neutral backgrounds. Generic models may introduce overly stylized or cluttered backdrops. Cloudinary’s updated model defaults to simpler, plainer settings that match retail norms. That saves teams from redoing basic edits and keeps catalogs visually consistent.
Combined with Refiners and Auto Enhance, this creates an end-to-end pipeline: AI generates the initial image, Refiners clean it up, provenance tracks its journey, and optimized delivery ensures it loads quickly on every device.
Refiners, C2PA, the Editor, Auto Enhance, and the updated GenAI model are available today on every Cloudinary plan, including the free tier.
To see them in action before rolling them into production, check out Clara Denari, Cloudinary’s transformation lab. It’s a lightweight playground where developers can try refinements, provenance, and automation in real time.
With this launch, Cloudinary is extending its core image platform to help teams not just manage and deliver assets, but also make GenAI practical in production. The aim is simple: faster workflows, authentic outputs, and visuals that consistently meet brand standards.
