The NessGate Charter

Version 1.0 · Our binding public promises. These will not be quietly walked back; any change is announced in the changelog.

NessGate exists to answer one small question, and to answer it neutrally:

Which machine-readable resources does this domain
itself publish right now?

What NessGate is

The smallest possible compatibility resolver for the AI web. Given a domain, NessGate reads whatever that domain already publishes across the standard machine-discovery locations (a data file, an API, an A2A agent card, an llms.txt, an ARD catalog, or a format that does not exist yet) and returns one normalized answer, with a link back to each source on the domain. It reads the domain on demand and caches each answer at the edge for up to 10 minutes, and stores nothing. Anyone (an AI agent, a tool, a person) can ask NessGate what a domain publishes, instead of guessing among unofficial pages, third-party databases, and scraped sources.

What NessGate is not

NessGate is not a catalog, a search engine, a crawler, a directory, or a data store. It does not rank, score, rate, or describe anyone. It does not host your content or define your format, and it makes no ownership, accuracy, or safety claim about what it finds. It is not a competitor to the resource formats and discovery standards below — it reads them and points back at them.

It works with everything

NessGate is deliberately format-neutral. When it reads a domain it recognizes, among others:

NessGate does not need to understand any of these deeply. It reads what the domain publishes and normalizes it. That makes it complementary to the emerging discovery ecosystem, not a rival to it.

Our promises

  1. Free forever to participate. There is never a fee to use the resolver, to be read, or to build on it. Free is not a launch promotion; it is the design.
  2. Never sell ranking, placement, priority, or badges. There is nothing to buy that changes how a domain is treated, because there is no ranking to buy.
  3. Neutral to everyone. Every domain is read the same way and returned the same way. No favored partners, no preferential treatment, no pay-to-win.
  4. The resource always lives on the domain. Your resources stay on your infrastructure, under your control. NessGate only reads them on demand and links back — it never stores, rehosts, owns, dictates, or becomes the resource.
  5. Open and independently implementable. The specification is public and the reference implementation is open source (Apache-2.0). Anyone may build a compatible resolver without asking permission. If nessgate.com disappeared, every resource would still stand on its own domain, unchanged.
  6. No accounts, no tracking. There is nothing to sign up for. NessGate keeps no user data and no domain data; this site sets no cookies and runs no analytics.
  7. A stable, additive specification. A response shape published for one major version keeps working across every later release in that version. A breaking change would only ever arrive as a new major version, with the old one supported through a long transition.

Honesty

NessGate is not the first to see this need, and does not claim to be. A whole family of open, neutral discovery infrastructure is emerging at once — among them Agentic Resource Discovery, AI Catalog, DNS-AID, MCP, A2A, and IETF drafts for domain-anchored machine discovery. We treat that as validation, not competition: it confirms the AI web needs a trustworthy way to read what a domain publishes. NessGate's contribution is to be the smallest, most neutral version of that primitive — one that takes no position on your format and asks nothing of the ecosystems it complements. We will never claim adoption, endorsement, or use that we do not actually have.

Read the technical details in the NessGate Resolver Specification, or learn more about the project.