Skip to main content

Love the app, concerned about privacy

Feedback

Hey I just love the app and it’s abilities! However, I’m concerned about privacy. Can you give some details about data handling etc.? Because I can see so much more use if more private documents are added.

Many thanks

Status: Completed3 comments

Log in to comment and vote

Comments3

  • Ryan

    Team•

    Jul 17, 2025

    Thank you so much for your feedback and for sharing your privacy concerns. We take privacy very seriously and are committed to protecting your data. You can find detailed information about our data handling practices in our Privacy Policy here: https://www.me.bot/privacy.

    If you have any further questions or need clarification, please don’t hesitate to ask!

    • Coffee Cactus

      •

      Jul 17, 2025

      I’ve been reviewing your privacy policy and was really intrigued by your mention of Confidential Computing, particularly the use of homomorphic encryption, data blurring and masking, and Azure Confidential Computing.

      I’m currently exploring solutions that prioritize advanced data privacy and security, and would love a bit more detail on the technical side of what you’ve implemented. Specifically:

      • Could you share which homomorphic encryption techniques you’re using (e.g. partially or fully homomorphic, and what operations are supported)?

      • What exactly is meant by “blurring” in this context — is it a UI-layer metaphor, or a data-level transformation?

      • Are your computations run inside Azure Trusted Execution Environments (e.g. Intel SGX enclaves), and if so, is that infrastructure auditable or documented?

      • When you mention complete anonymity, is this achieved through true anonymization, pseudonymization, or zero-knowledge techniques?

      • Lastly, is there any available technical whitepaper or security audit that goes into more depth?

      I appreciate the privacy-forward posture you’re aiming for — just hoping to better understand how it's actually implemented. Looking forward to your response!

      • Ryan

        Team•

        Jul 17, 2025

        Hi Tobes,

        Great questions. Here’s a more direct overview of our approach to each of the areas you mentioned:

        1. Homomorphic Encryption: We primarily utilize Partially Homomorphic Encryption (PHE) as a key component of our privacy-preserving computation stack.

        We use it to perform specific calculations, like searches, directly on encrypted data. This allows us to provide you with a useful service without ever needing to see your actual content.

        2. Data "Blurring":

        This is indeed a data-level transformation technique, not a UI metaphor. Internally, we refer to this as applying Differential Privacy.

        Before we analyze data for group statistics, we add a layer of mathematical noise. This makes it impossible to identify or reverse-engineer any single person's information from the results, protecting your individual privacy within the group.

        3. Azure TEEs: Yes, we use Azure’s Confidential Computing. Think of it as a secure "black box" (a TEE or enclave) on our servers where authorized computations happen. This environment is encrypted and isolated, meaning even we can't access the data being processed inside. It’s a powerful second layer of security on top of the primary encryption controlled by your local key.

        4. How We Achieve Anonymity: Our approach is built on user control, combining a few key elements:

        • Your Key, Your Data: The key on your device is required to decrypt your data. We don't have it.

        • Your Permission: Sensitive actions like data exports require your explicit approval through a Threshold Signature. We can't act without your device's participation.

        • Anonymous Analytics: For any group insights, we use Differential Privacy to ensure the results are fully anonymized.

        5. Technical Whitepaper: Absolutely. A technical whitepaper covering our architecture is in the works. We'd be happy to share it with you when it's ready. If you're interested, please contact us at support@mindverse.ai so we can notify you.