Roll The Credits

Roll The Credits combines traditional steganography techniques with deep learning to securely hide sensitive information (like crypto wallet recovery phrases) in media files. The project offers multiple encoding methods, from simple image manipulation to advanced neural network-based approaches.

Our philosophical vision - we keep secrets to protect facts, truths that should be accessed by some, and not by others.

Have you ever had the need to look over your shoulder while managing your private keys? Our current solutions can stop a crypto-burglar from stealing your private keys due to unintentional visual exposure - potentially ruining your financial future.

What can we do for you? Roll The Credits transforms the paradigm of privacy by hiding encrypted or protected data data in plain sight using deep steganography with additional layers of security. Think of it as a safe capsule on top of cryptography, we believe steganography complements cryptography by abstracting existence of encrypted data in the first place. Our application allows to embed free input data or custom mnemonic phrases into Cannes movie posters or even better - credit scenes of famous movies, where few would look.

To illustrate the essence of new emerging deep steganography, consider a simple example: the phrase “The weather today is quite pleasant” could be transformed to “The climate today is rather nice” to encode the binary message “1011”. Each word substitution (weather → climate, quite → rather, pleasant → nice) represents a specific bit pattern, invisible to casual readers but recoverable with the correct key. Deep stego enhances this approach by using GPT-4-omni to generate contextually appropriate synonyms that maintain natural language flow while maximizing encoding capacity.

On top of that user can embed additional security layers such as:

• Keeping data secure even if exposed with password protection.
• Add duress stress mode to display decoy data in case wrong password is provided.
• Request World ID or similar proof before data is reveled.
• Request on-chain payment proof before data is reveled.
• Encrypt encapsulated data with additional private key.
• Allow data to be displayed only once or deleted upon unauthorized access.

Hidden in plain sight, yet avoids screenshot snipes & sensitive data accidental visual exposure during transfer. Our app also supports multilingual and custom wordlists, disguise your seed as a mix-tape or movies.

We built this project using Hardhat, Node.js, React, and a steganography algorithm, each playing a crucial role in delivering a decentralized and creative solution.

At the core, Hardhat was our Ethereum development environment - it allowed us to write, test, and deploy smart contracts quickly. We used it to manage contract compilation, local blockchain simulation, and deployment scripts. On top of that this helped us to generate private keys for our use case before embedding them into Cannes movie posters. We've also created additional MCP server for Hardhat 3.0 for the convenience of using it as simple as chatting with an agent.

On the backend, Node.js served as the foundation for our API layer. It handled off-chain logic and acted as the bridge between our front-end interface and the blockchain. We used Express.js to expose endpoints that could encode or decode data using our steganographic techniques.

The front end was built with React, which gave us a fast, interactive user experience. We used web3.js (or ethers.js, depending on the environment) to connect the front end to deployed contracts and allow users to interact with the blockchain directly in their browser.

Now, for the hack: we used a steganography algorithm to hide data - specifically secret messages or hashes - inside images and credit scenes (video). These were published on platforms like Medium, which allowed us to effectively store off-chain secrets in plain sight. This added a unique, censorship-resistant layer to our system. The images look like ordinary media identical to original source, but they contain embedded, encoded data that our app can extract and verify.

This approach allowed us to demonstrate essence of steganography and explore ideas of Secret Images if these were options for popular wallets such as MetaMask or RABBY-Wallet to allow users migrate, import or export their private keys without revealing mnemonic phrase. This would minimize unwanted exposure and protection from clipboard scanners or malicious actors trying to snipe your private keys during creation, import, export or migration between wallets.