Why AIWebIndex is an open protocol.
The web only works because its base layer (HTML, HTTP, the robots.txt convention, RSS, schema.org) is open. The way AI systems read the web should be answered the same way: with a shared protocol, not a captive product.
Protocol versus product
AIWebIndex is a protocol: a User-Agent identifier, a JSON document format (AIDocument), and a verification mechanism. It is unpatented and freely implementable. Anyone can build a crawler, an indexer, or a verifier that conforms to it; the conformant ones interoperate.
Lyrenth, run by Aleksma AI Inc., is a product: a hosted API + dashboards that implement the protocol commercially. Lyrenth is the reference implementation: useful as a working example, useful as a paid option for teams who want hosted infrastructure rather than running their own. It is one implementation. Not the only valid one.
This site exists to make that distinction visible. If AIWebIndex only existed inside Lyrenth, the “open protocol” claim would be marketing copy. The protocol has to live in a place where Lyrenth does not control the editor pen. That place is here.
The pledge
Aleksma AI Inc. (the steward) commits, formally:
Aleksma AI Inc. holds no patents on the AIWebIndex protocol’s core mechanics (the User-Agent identifier, the AIDocument format, the verification mechanism) and pledges not to seek such patents. Implementations are free to build, fork, and extend. Aleksma AI Inc. will not assert intellectual-property claims against implementations that conform to the protocol or that derive new protocols from it.
This pledge is binding for 1.0 and any future versions Aleksma AI Inc.publishes here. If a future version of the protocol is forked under different stewardship, the new steward’s pledge governs that fork.
Why open standards win
Search runs on robots.txt and sitemap.xml because they are open. Email runs on SMTP, IMAP, DKIM, and SPF because they are open. Identity runs on OAuth and OpenID Connect because they are open. Structured-data on the web runs on schema.org because it is open. None of these are perfect specifications. All of them won, decisively, against closed alternatives.
The pattern is consistent: when a layer of the web becomes shared infrastructure, it has to be a protocol that anyone can reimplement. Closed alternatives at infrastructure layers consistently lose to open ones, even when the closed version has a head start in tooling, marketing, or distribution. The history is well-attested.
AIWebIndexis making an early bet that the layer of “how AI systems read web pages” will be the next instance of this pattern. Lyrenth earns long-term durability by stewarding the layer well, not by owning it.
Governance & revisions
1.0 is the first published draft. Future revisions follow a small process:
- Proposed changes are discussed publicly via email to hello@aiwebindex.org and on the GitHub repository.
- Implementer feedback is treated as primary input. If the reference implementation and at least one independent implementation both signal that a change is needed, the proposal advances.
- Backwards-compatible additions land in minor versions (1.0 → 1.1). Breaking changes require a new major version (1.x → 2.0) with a documented migration path.
- Aleksma AI Inc. retains editorial discretion only over the published version. Anyone is free to fork the spec and steward an alternative.
Contact
For protocol questions, implementation questions, listing requests, and anything else: email hello@aiwebindex.org. For commercial inquiries about the Lyrenth implementation specifically, the Lyrenth site has its own contact channels.