Capabilities and Permissions

The Hypermedia protocol brings a permission system to the web, building upon our identity system, and robust cryptographic primitives. Our capability system is inspired by UCANs Capabilities Permissions

Open Editing

In the web, there is no widespread protocol for collaborative editing. And there is no mechanism to create "your version" of some other web page, while keeping authorship of the previous version. This constrains collaboration between people who don't yet know each-other. The Hypermedia protocol allows collaborative editing of documents, between identities who have been granted permissions. We also support "Open Editing" which allows you to make changes to documents that you don't have write permissions for (as long as you publish your version on your site or another site where you have permissions). Our open editing model is largely inspired by Git- the world's most successful distributed version control system. Collaborative Editing Document Branching

Version History

On the web, we have no way to track changes to content over time. As a result, we have to build custom applications to support version history, for example Wikipedia (MediaWiki). As a result, if you want to see the version history of a website, you must rely on the archives from a trusted third party, such as Archive.org (the "wayback machine"). With lack of protocol support, there is no browser support for traversing the history of web pages, or to see the authorship of each change to the document. The Hypermedia protocol brings version history and authorship to the web. Document Change History Comment History

Deep References and Embedding

The traditional web allows you to form links- but there is no reliable way to link to specific content within a page. And there is no way to link to an older version of a document on the web. The web also suffers from a lack of native embedding (aside from "iframe", with it's ergonomic issues and lack of target document addressing). Fine-Grained Links Versioned Links Embeds Combined with the Community Archival feature of the protocol, we empower knowledge managers with a very robust system of knowledge preservation and cross-site collaboration.

Community Archival

While the traditional web depends on a small number of organizations for archival, the Hypermedia protocol natively supports archival. You can access and store the source data for the content you view, and redistribute it, if you want. Because we build upon IPFS, and because all authors sign content, you don't need to trust the archiver to give you valid data. You can validate it yourself by checking signatures and the content-addressed IDs (CIDs). The End of Broken Links

P2P API

Peers in the Hypermedia network connect over libp2p, and they exchange gRPC messages. Primarily, the peers are exchanging references to  IPFS blobs. These blobs are then exchanged with the IPFS bitswap protocol, and are then validated as a part of the peer's local indexing process. This API is not yet stable and remains undocumented. To learn more, you refer to the protobuf definition files:

Signed Data

Agent Skill

We have created a simple Agent Skill which you can use with a wide variety of LLM Agents: Seed Hypermedia Agent Skill To install the skill on your machine, run: npx skills add https://github.com/ericvicenti/seed-hypermedia-skills --skill seed-hypermedia

CLI

You can connect to the HTTPS API of a Hypermedia web site, using the Seed CLI:

JSON-REST API

We have a JSON REST API which allows developers to access a Seed node who is exposing a web server. This is used by the front-end of any Seed web site. Not currently documented. Coming soon to the protocol documentation. You may browse the code entry point here on github, and the endpoints are defined here.

Hypermedia Explorer

You can browse the raw Hypermedia data using explore.hyper.media. This app is built by the Seed team, and utilizes the HTTPS API on the hyper.media host. It serves to help you understand and debug the Hypermedia protocol.

Seed Web

The reference implementation of Seed Hypermedia for web. This shares most of its code with . On Seed Web, the P2P Node runs on the server-side. Clients rely on the server for data integrity, but the user's private key is not accessible to the server. You can set up a Seed Web site through the interface in the Seed Desktop app, and you are encouraged to set up a custom domain. With a private key, backed up content on the desktop app, and a custom domain, you can retain control over your online presence, even if something bad happens to the Seed servers. This web server exposes an HTTPS API which may be used by clients who are unable or not willing to access Hypermedia content directly via the P2P API.

Seed Desktop

The flagship software developed by the Seed team, the Desktop app acts as a browser, archiving tool, publishing tool, and P2P server. You may download the app for Mac, Windows, or Linux: