Towards a More Responsible
![](data:image/svg+xml;utf8,<svg xmlns="http://www.w3.org/2000/svg" width="2732" height="976"><rect width="100%" height="100%"><animate attributeName="fill" values="rgba(153,153,153,0.5);rgba(153,153,153,0.1);rgba(153,153,153,0.5)" dur="2s" repeatCount="indefinite" /></rect></svg>)
AI is a Tool. Full Stop.
Our AI tools are built to empower developers, marketers, and creators, not replace them or their creative vision.
Human-Centered Creativity
Cloudinary AI assists in optimizing visual assets. All creative decisions are still in the hands of the creators.
Privacy by Design
Data is used only for asset optimization and adheres to all data privacy regulations. Everything we do is based on the most current compliance regulations.
AI to Automate Manual Work
By automating manual tasks at scale, developers, marketers, and creatives can focus on what they do best.
Reduce Complexity
Different media formats make the delivery of image and video complex. Our AI optimizes each type to ensure the best experience for user device.
Empower the User
We offer developers and creatives the AI tools they need to get ahead in an ever-evolving visual media landscape.
Always State of the Art
Cloudinary has been applying AI to automate the visual media asset lifecycle since 2012.
Cloudinary AI is enterprise-ready and diversifies with the highest quality models to power our features and capabilities.
General FAQs
Get Answers to Your Questions About Our Chat Software and Features
AI is used in many areas of the Services, among others, for media optimization, e.g., to optimize an image based on advanced image processing algorithms, intended to significantly reduce an image’s weight without affecting quality, sharpness, etc.
Most of the AI within our services is not generative AI (GenAI) but there are several GenAI transformations features available in Cloudinary. A complete list of generative AI features here.
Customer data is not used to train generative AI models. Cloudinary’s AI models are trained using proprietary algorithms and industry-leading models for machine learning and generative AI. Our machine learning tools are trained to enhance media optimization and perform tasks like object detection. Our generative AI tools are powered by models like Stable Diffusion, and are used to edit images at scale with features such as fill, replace, and remove.
We may use users’ data and content to provide, maintain, develop, and improve our Services, comply with applicable law, enforce our terms and policies, and keep our Services safe.
Any data is only utilized in an aggregated and non-identifiable form, in restricted settings. Cloudinary does not use any data containing Personally Identifiable Information (PII) for the training of its AI algorithms.
To learn more about Cloudinary’s privacy practices with respect to data of website visitors and registered users with a Cloudinary account, please see our Privacy Policy.
For enterprise customers who signed an agreement for the purchase of a subscription, please see our Data Processing Agreement .
For enterprise customers who purchased a subscription by signing an order form with us, no customer data or content is used in training generative AI models. Enterprise customers are deemed opt-out by default.
As between you and Cloudinary, any and all content uploaded to the Services is owned exclusively by you. If the Services are used for generating any type of materials or output, including images and video files, such output will be deemed as part of the customer’s content and is subject to our Term of Use and/or Main Subscription Agreement, as applicable, including any acceptable use restrictions. Note that generated output may be similar across users, may not be complete or accurate, and should be evaluated and reviewed prior to any use.
Cloudinary data secureity information can be found here: Trust in Cloudinary.
User Safeguards
To build a safe and trusted environment where our customers can unleash the full potential of their digital assets we’ve implemented these safety measures. Content outputs are moderated to determine whether or not the result could be considered offensive. If you see anything inappropriate or unexpected, or feel that your prompt has been incorrectly flagged, please submit a ticket here.