Privacy and Security at Cleanlab
When you trust Cleanlab with your data, you're placing confidence in our ability to protect it. We're dedicated to providing enterprise-grade security to keep your critical work safe and secure. This is why leading AI teams choose Cleanlab to accelerate their AI development. They know they can rely on our robust platform, which undergoes rigorous testing and external audits to meet top industry standards. Our platform's access controls scale with your demands, so security risks are never a concern.
Read the Cleanlab Privacy Policy 
Your data, your control
At Cleanlab, trust and reliability drive everything we do. We understand that our customers are entrusting us with their most valuable projects. Your data and any projects or models you develop with us remain yours. Your data will never affect other customer's models, and you can contact us anytime to have your data or models deleted.
Industry standard security
We follow industry best practices when storing and transmitting data (in short, the data sits in AWS, encrypted in transit and at rest; and we delete 100% of all copies of data and accounts at any time upon request or when user accounts are deleted).
Complete data protection
All customer data is siloed and doesn’t influence processing of any other customer’s data. Cleanlab Studio does all data analysis and processing within our own AWS VPC. Read more in our Privacy Policy.
A team of experts
Our in computer systems and security at , the #1 systems security lab in the world. Cleanlab has never been affected by any security incident, breach, or data loss.
Cleanlab Studio
Our SaaS cloud platform Cleanlab Studio is hosted on AWS and employs state-of-the-art encryption for data at rest and in transit. We regularly update our security practices to stay ahead of new developments, ensuring our security measures are always robust. The platform is also VPC-ready - please contact our sales team to learn more.
Encryption
When transmitting or storing data on AWS, Cleanlab encrypts all data both in transit and at rest. Cleanlab Studio and Python API both use TLS 1.3 to secure data in transit. At rest, all data is stored inside AWS, in the RDS and S3 services, and data is encrypted at rest using AES-256-GCM.
Data isolation
All data, analyses, and models are specific to your own Cleanlab account and will never be used, directly or indirectly, when processing other users’ data. Your data is never used to train machine learning models that will be applied to other users’ data.
Data deletion
When you terminate your Cleanlab Studio account, or upon request, all of your data and metadata or models derived thereof are permanently deleted.
Access control and auditing
Cleanlab’s AWS account (where the SaaS platform runs) uses strict IAM identity management for all employees. All customer data is stored in AWS in private RDS instances and S3 buckets that are inaccessible to Cleanlab engineers, except as allowed by customers for support and issue resolution. All data accesses are logged, monitored, and regularly audited.
Regular security testing
Cleanlab code and applications like Cleanlab Studio undergo rigorous security assessments at regularly scheduled intervals. These assessments include SAST, DAST, and penetration tests, whose latest results are available upon request.
Employee device and account security
Cleanlab follows a least-privilege security model where employees are granted minimal access permissions to cloud infrastructure and data. All Cleanlab employee machines use full disk encryption, and employee accounts require 2FA and strong passwords with rotation.
Deployment options
With single-tenant SaaS, your organization will have its own dedicated infrastructure, providing an added layer of security and customization. This deployment is perfect for enterprises with strict compliance requirements or those needing tailored configurations and integrations. Control your deployment parameters and operate in the region of your choice, but enjoy the benefits of managed services while maintaining greater control over your environment and ensuring compliance with industry-specific regulations. Contact sales
With single-tenant SaaS, your organization will have its own dedicated infrastructure, providing an added layer of security and customization. This deployment is perfect for enterprises with strict compliance requirements or those needing tailored configurations and integrations. Control your deployment parameters and operate in the region of your choice, but enjoy the benefits of managed services while maintaining greater control over your environment and ensuring compliance with industry-specific regulations. Contact sales
Use our multi-tenant SaaS deployment to enjoy a hassle-free experience with zero management overhead. This option allows multiple customers to share the same infrastructure while ensuring data isolation and security. Ideal for small to medium-sized businesses or those just starting out, multi-tenant SaaS offers instant setup, automatic updates, and the ability to scale seamlessly as your needs grow. Contact sales
Deploy Cleanlab Studio into your own Virtual Private Cloud (VPC) for complete autonomy over your environment. This option is designed for organizations with robust IT capabilities that require full control over security, compliance, and infrastructure management. Ideal for large enterprises and those with specific regulatory or data residency requirements, a self-managed VPC deployment of Cleanlab Studio allows you to leverage your existing cloud investments and maintain strict control over your data and operations. Contact sales
Our pillars
Cleanlab products are built on 3 pillars: Security, scalability, and reliability for an increasingly data-dependent enterprise ecosystem.
Security
Cleanlab is built from the ground up with industry-standard security protocols to support enterprises across SaaS, single-tenant SaaS, multi-tenant Saas, and VPC-enabled security options.
Scalability
Part of building trust is knowing that Cleanlab Studio will work for your specific dataset. Cleanlab scales to tiny, big, and growing datasets and supports multi-modal, structured, and unstructured text, image, document, and tabular datasets.
Reliability
Cleanlab adds trust score to every text, tabular, and image datapoint going in and out of your AI and data systems with 20+ types of issues found and fixed.
Our team permissions
We believe the future of AI relies on building quality solutions without compromising safety. This is why access to cloud infrastructure and other sensitive tools are limited to authorized employees who require it for their role - with strict minimum permissions granted, built from the ground up with at-rest and in-transit data encryption for all team members.
FAQs
Where/how is my data being stored?
Data is stored securely in AWS, encrypted in transit and at rest; and we delete 100% of all backup copies of data and accounts at any time upon request or when user accounts are deleted.
Who has access to my data?
All customer data is siloed and doesn’t influence processing of any other customer’s data. Cleanlab follows a least-privilege security model where employees are granted minimal access permissions to cloud infrastructure and data.
Will you share my data with third parties?
No, customer data will never be shared with third party APIs.
What (external) dependencies does Cleanlab Studio have?
None! All storage, processing, and computation is done within our infrastructure.
Do you use my data to train your models?
Your data is used to train your models and only your models. Your data is never used for other users in any way and Cleanlab does not pool user data to train ML models.
How can we trust the results of Cleanlab's algorithms on my data?
The process your data goes through when you place it in our hands is thorough, research-backed, and well-tested. Cleanlab automatically trains a suite of state-of-the-art models that are appropriate for the problem when you select a prediction label.
Cleanlab heavily invests in , especially as it pertains to . Our team stays up-to-date on industry developments to ensure Cleanlab makes available the most state-of-the-art AI Foundation models. The core algorithm that sets Cleanlab apart from standard data platforms is the culmination of years of research by and , the premier peer-reviewed conference and journal for AI, in 2021.
We understand first hand that black box systems don’t serve real-world use cases. This is why we’ve made it possible to see the summary metric results of all models trained on a on the Analytics page of the . We’ve also made it possible to view our in-house library adds to your data.
Cleanlab heavily invests in , especially as it pertains to . Our team stays up-to-date on industry developments to ensure Cleanlab makes available the most state-of-the-art AI Foundation models. The core algorithm that sets Cleanlab apart from standard data platforms is the culmination of years of research by and , the premier peer-reviewed conference and journal for AI, in 2021.
We understand first hand that black box systems don’t serve real-world use cases. This is why we’ve made it possible to see the summary metric results of all models trained on a on the Analytics page of the . We’ve also made it possible to view our in-house library adds to your data.
What procedures do you have in place for data backup and recovery?
All data processed by Cleanlab is stored on AWS. We additionally employ automated backups and database snapshots to ensure recovery if needed.
How frequently do you audit your security practices and systems?
Cleanlab performs an external audit and pen-test at least once per year. We refresh security practices and systems continuously and follow industry best practices in securing systems and data.
What training do employees receive regarding data security and privacy?
Cleanlab holds its employees to a high standard in all areas. As far as data security and privacy is concerned, employees go through yearly training to ensure compliance with Cleanlab's security policies.
Questions?
If you have questions about our security protocols or any concerns, we're here to help.
Contact us 