Enthusiasts have found a way to use YouTube as cloud storage for any files

A project has appeared on the GitHub platform that offers a way to store files on YouTube as videos without losing data. The development, called yt-media-storage, actually turns video hosting into a storage, using the features of processing and uploading content to the service.

Encoding files into a lossless video stream

The essence of the solution is to convert arbitrary files into a lossless FFV1 video stream packed into MKV containers. After that, the resulting videos can be uploaded to YouTube as a regular video. If necessary, the data is extracted back and restored to the original file without changes. Thus, the author of the project proposes to use free and formally unlimited video upload as an alternative to traditional cloud storage services.

The tool was developed by a GitHub user under the nickname PulseBeat02. The project is distributed as an application with a command line interface, as well as a graphical shell designed for encoding and extracting data. It is based on a multi-level redundancy architecture designed to increase storage stability. To correct possible distortions that occur during video processing on the YouTube side, Wirehair fountain codes are used, which allow you to restore lost fragments.

Before conversion, the file is divided into parts that are encoded using Wirehair, and then embedded in video frames at the stage of video formation. When decoding, the system extracts packets from the video and assembles the original file. Additionally, encryption is implemented using the libsodium library and the XChaCha20-Poly1305 algorithm, which provides data protection before uploading to the platform.

The encoding process is carried out in 4K 3840×2160 resolution at 30 frames per second. The developer has provided redundancy mechanisms designed for the specifics of the YouTube video pipeline. The FFV1 codec refers to lossless intra-encoder compression formats and, unlike solutions such as HEVC. This allows you to save every bit of the original data, and then completely restore the file.

Technical requirements and methods of use

The project supports several interaction options. Users have access to a command line interface and a graphical program. The terminal allows you to run the encoding using commands, while the graphical version allows you to work in a more familiar environment and supports batch processing – several files can be queued for sequential conversion.

For compilation, CMake version 3.22, a compiler with support for the C++23 standard, FFmpeg, libsodium, OpenMP and Qt6 with the Core and Widgets modules are required. If the FFV1 encoder is to be used in the mp4 container, FFmpeg version 8 or later is required. The user also needs to be able to compile the source code project in C++23. At the same time, the author provides ready-made compilation artifacts for CLI and GUI via the CI/CD pipeline.

Despite the technical development, the project has already caused a discussion of the legal aspects of its use. On the Hacker News discussion site, participants drew attention to a possible violation of YouTube’s terms of use and the risks of account blocking. The service’s rules explicitly state that the platform is intended for viewing and sharing video materials, and not for working as a universal cloud file storage.

According to the GitHub repository, the project has collected several hundred “star” marks, dozens of forks, and a number of commits in the main branch, which indicates attention from the community and development. At the same time, there are concerns in the comments that the use of such an approach could lead to sanctions from the platform. One of the discussion participants noted that in such cases, users are more likely to ask about the timing of a possible blocking than about the probability of such an outcome.