Hi, it is Ultranymous. We are developing interplanetary storage for scholarly works, [Standard Template Construct](https://www.reddit.com/r/scihub/comments/12detqs/standard_template_construct_store_and_search_the/) (STC).
https://preview.redd.it/5extcj95c8wa1.png?width=914&format=png&auto=webp&v=enabled&s=8bde9ea40b66fdd3d07beebb378457228d916baf
Now, we have built programmatic tools to access the STC database through API. This API allows you to perform search queries and retrieve metadata for papers, including recognized full-text content for some of papers.
The metadata may also contain a unique CID that allows you to download selected papers through IPFS. Please note that not all records are available yet, as we are still in the process of uploading collected papers. Our priority is to upload papers that are absent in other libraries, so it is worth using [STC](http://standard-template-construct.org) that also provides ready cross-links to Sci-Hub, LibGen and other sites.
Like STC, this API works over IPFS and has its pros and cons. The pros include
* **total decentralization,** which means that nobody can ban you from using it and nobody can ban us from providing the service.
* no limits on downloading; no CAPTCHAs, no rate-limiters, nothing! This makes it well-suited for mass-processing of papers or for building accessing services like Sci-Hub or Telegram bots on top of STC API
However, the cons include slower service, which means that it is not suitable for millisecond apps. We recommend using it for offline processing or cherry-picking articles for local use. Multiseconds requests are usual, despite aggressive caching of everything.
**How to use?**
To use STC API, you should have [IPFS](https://docs.ipfs.tech/install/ipfs-desktop/) and [Python3](https://www.python.org/downloads/) installed.
Then, install our toolkit for accessing STC by running the command:
pip3 install stc-tools
It includes a CLI for performing searches and downloads, for example (first query will be slow, then faster):
stc-tools – download doi:10.1177/1745691612459058 file.pdf
Additionally, toolkit contains a small utility class for use in your Python apps. You can check out an example on our [GitHub](https://github.com/nexus-stc/stc/blob/master/tools/examples/search-stc.ipynb) repository. You can also use it to report bugs and provide feedback.
**How to help?**
* We are looking for people to participate in building our IPFS Cluster, especially those with good bandwidth, 5TB+ of free disk space, and 16GB+ RAM. Participating is a direct way to serve as pillars and seeders for humankind scholarly knowledge across the planet
* We are also seeking librarians with access to various libraries to participate in our sharing collective, donating small to receive a lot
* You can follow the [STC guide](http://standard-template-construct.org.ipns.localhost:8080/#/replication) and pin some part of the collection, which does not require interaction with us. We may preserve our knowledge only if a lot of people participate, and we have made participation easier
* You can also try building any service on top of the STC API. For instance, you may implement a Telegram bot that delivers papers or chain STC with Large Language Models to create a service for answering scientific questions
* STC requires a large high-memory, high-disk, and high-CPU server for its operations. If you are ready to donate such a server, you will get a complete and OCRed collection of scholarly texts much sooner. Just imagine what is possible if GPT models meet the entire corpus of science
* We invite you to participate in our mission. Tell others about us, join our [channels](https://t.me/nexus_search), and spread the word. We need people to learn that knowledge is breaking its chains. Words are fuel for our brains, and knowledge is the core of our civilization. Given the presence of quite strong AI that is working well with texts, scholarly written knowledge starts to become a doomsday weapon and hence knowledge owning becomes owning of the weapon. It must be liberated from all legal barriers as soon as possible to make a competition in creation of researching AGI fair and equal for all.
You may reach me at [ultranymous@proton.me](mailto:ultranymous@proton.me) for discussing any of the points above.
https://preview.redd.it/jeqeldc8c8wa1.png?width=686&format=png&auto=webp&v=enabled&s=feabc9bb804f420efb66e763ec54575103097085