How our security measures work

Steganography, like cryptography, is a technique within the field of information security, focused on protecting data (secret) by hiding it within another medium, such as an image or file. To achieve this, the medium is manipulated in a way that makes the embedded secret undetectable to anyone with access to the altered medium. A common example is modifying the least significant bits of an image’s pixels to hide “identifiers” for purposes like fingerprinting or watermarking. The primary goal of steganography is to keep the secret undetectable, with algorithms dedicated to identifying areas in the medium that can be subtly altered to achieve maximum invisibility. While these alterations may remain invisible to the human eye, advancements in computer vision and AI have made detecting such modifications increasingly feasible, requiring resource-intensive AI algorithms for effective hiding the secret against these attacks.

At Secrets Vault, we deliberately chose not to manipulate images to hide information, and therefore, our approach cannot be properly classified as steganography. Instead, we follow a visual cryptography approach—specifically, a secret-sharing scheme. In our method, images are not used as a container for hidden information, as in steganography. Rather, an image (the ‘Keepic’) serves as one of the essential shares in a cryptographic secret-sharing protocol used for protecting a secret.

One common way to protect sensitive information is by encrypting it with keys created from passwords. This method, called Password-Based Key Derivation, is used to secure things like files or digital certificates. Typically, a secure key length of at least 128 bits is recommended. But in encryption, it’s not just about the key length; it’s also about how strong the password is that generates the key. For example, if a key has 128 bits but is generated from a password with only 50 bits of strength, the security is limited to 50 bits, making it easier to break. This is why passwords need to be both long and complex to provide the best protection—ideally, they would reach a security level close to 128 bits.

However, making passwords secure enough often means they become harder to remember. For example, using uppercase and lowercase letters, numbers, and a few special characters, each character adds about 6 bits of strength. To reach 128 bits, a password needs at least 21 characters, which results in something difficult to memorize, like ‘jlE5FSPTT7fi4YnZ63Xy2.’

Could an image serve as an alternative? In terms of security, an image is made up of many pixels, with each pixel containing color information. Even a small 200 x 200 image, for example, can provide up to 120,000 bits of potential security—nearly 1,000 times the security of a long password. This shows that images could be a strong alternative to long, complex passwords.

Yes, as long as alterations don’t significantly change the core of the image (e.g., resizing rather than cropping). Our protocol is designed to withstand common image manipulations encountered when storing or sharing images on external platforms like social networks.

Unlike other methods, where even a single pixel change can prevent recovery, Secrets Vault’s technology processes the image to extract unique, invariant information. This ensures that typical modifications like compression, resizing, or watermarking won’t hinder recovery. This approach makes storing and sharing your Keepic more flexible and secure than encrypting it directly with a derived key.

No, our cryptographic scheme has no specific limitations on the type or length of the secret. However, technical constraints on the client device performing the cryptographic operations may apply. Currently, we limit the secret size to 8 megabytes for browser clients, but we aim to increase this to around 100 megabytes soon.

The only limitation is on the minimum image size to ensure sufficient bits can be extracted for secure protection. The current lower limit is 50,000 pixels (approximately a 224×224-pixel image). While we may reduce this limit in the future, it’s already low enough given that most images used today exceed this size. For example, a VGA-resolution image (640×480 pixels) has over 300,000 pixels.

No. Our scheme is designed to maintain its security strength regardless of the image size, allowing you to resize the original image (Keepic) without compromising security. We currently set the minimum resize limit at 50,000 pixels (e.g., 224×224 pixels) to ensure 256-bit quantum-resilient security. We are working to further enhance and potentially reduce this limitation in future versions.

The Keepic is essential for recovering your secret; without it, recovery is impossible even with access to the Kvault. Losing the Keepic is like losing the keys to an encrypted secret. We recommend keeping several copies of the Keepic in different, reliable locations. Since the Keepic can be any public image, you could choose one that is widely accessible (e.g., a museum painting) or store it on public platforms or social media without concern for compression or resizing. Since the Keepic itself is unaltered, it’s indistinguishable from standard images, making it blend seamlessly among publicly available or shared images.

No, the Kvault alone isn’t enough to recover the secret. It’s created through a secret-sharing method that also requires the Keepic to complete the recovery. This makes Kvault safe to store with a trusted third party if needed. To protect your secret, it’s best to keep the Keepic and Kvault separate, as having both allows full recovery—similar to storing an encrypted file alongside its password. The Kvault on its own doesn’t reveal any information about the secret, so there is no risk of security compromise.

Currently, Secrets Vault also stores a Vault share needed for recovery, making the process easier for users. However, we’re developing a decentralized storage system with Multi-Party Computation, which will make storing and retrieving the Vault share even more secure and resistant to interference.

Our cryptographic scheme is designed to execute entirely on the user’s device, ensuring that neither the secret nor the Keepic is ever exposed to third parties. The only component that leaves the user device is the Kvault, which is securely custodized by the Secrets Vault platform. Along with the Kvault, a Keepic commitment is sent to Secrets Vault for future retrieval. This commitment enables Secrets Vault to validate a zero-knowledge proof provided by the user before releasing the Kvault, confirming that the user has the correct Keepic without revealing it. This setup ensures that the Keepic remains private and prevents impersonation attempts by unauthorized users.