Architecture
What the Human Web is.
Architecture as a position on what the internet should be for.
Definition
The Human Web is the part of the agentic web where people — not platforms — are the primary actors, and where four things stay with them: their identity, their agency, their self-representation, and their value. Together, these are what we mean when we say someone's "self" is held as their own.
It is a set of rules, standards, and commitments that a given piece of the internet either meets or doesn't. Most of today's internet is outside the Human Web. That's expected. The Human Web is not trying to replace the existing internet — it sits across it as a distinct layer of identity, agency, self-representation, and value, while the content, applications, and transactional surfaces of the existing internet continue to operate alongside.
What it is for
The Human Web exists for human flourishing — what people do and what they become.
Beneath that named purpose sit two paired architectural commitments: to human agency, and to the conditions for collective wellbeing. The first is what makes the Human Web human-centred at the level of the individual. The second is what makes it serious about the world those individuals share.
It is the digital infrastructure built around the proposition that the actions of a life — contributions to others, participation in communities, the labour of caring, the work of becoming — and the states of life that follow from them — health, capability, belonging, security, the sense of a life going well — are first-class concerns of the systems we build.
This is not a metaphysical claim. It is an observation about which kinds of digital activity the Human Web takes seriously. Generic enterprise automation, attention-extracting content platforms, and pure-transaction marketplaces operate in other parts of the internet; they are not refused, they are simply not what the Human Web is for.
The six entry conditions
A service, product, or system is inside the Human Web if it meets all six of the following entry conditions. Together they specify what it means, architecturally, to hold a person's identity, agency, self-representation, and value as their own.
- 01
Sovereign identity.
The user holds an identity that is theirs — a persistent identifier that no single service can delete, alongside pluggable authentication services they choose, under trust frameworks set by domains rather than imposed by individual providers. Identity is decoupled from authentication, which is decoupled from trust framework. This decoupling is what makes Human Web identity portable across services and across time.
- 02
Agent representation.
The user can act through agents that represent them — not through accounts inside a provider's UI. Agents hold the user's permissions, context, and preferences, and work for the user across systems, not inside any one of them.
- 03
Reflection.
The user has the means to see, interrogate, and shape how the system represents them. This is more than data access. It is the architectural commitment that the user's self-representation — the picture of them held inside the Web — is theirs to understand and theirs to correct. Reflection sits between agent representation and permissioned data because it is what makes the first meaningful and the second navigable: agents act for the user only if the user can see what those agents understand about them; consent over data is only real if the user can see what is being held and inferred.
- 04
Permissioned data.
Data flows in and out of services only through explicit, scoped, and revocable user consent. The user's data does not accumulate as a silent by-product of participation.
- 05
Verifiable contribution and value.
Verifiable facts about the user — their actions, contributions, achievements, and the attributes that have been certified about them — are recorded as credentials issued by trusted issuers. Credentials are portable, can be presented selectively, and are recognised across services. Value flows back to the user across system boundaries.
- 06
Portability.
The user can leave. Their identifier, credentials, agents, data, self-representation, and consent history can be moved to another Human Web-compliant service without loss.
Together, the six conditions specify what an orchestrator — an entity that builds and operates a Human Web-compliant Pass-and-agent system — has to do to be inside the Web rather than adjacent to it. They are the architecture's working definition of compliance.
The wellbeing economy connection
The Human Web's purpose connects to a longer conversation. The wellbeing-economy movement — twenty years of work by economists, civic bodies, and governments — has argued that GDP-centric measurement misses what matters and that economic activity should be evaluated by its contribution to human flourishing. The movement has had clear philosophy and patchy operations.
The Human Web is, substantively, the missing operational layer. Its rules ensure that people remain sovereign in the recognition and routing of their own lives. Its architecture supports the agents — on both user and organisation sides — that produce, recognise, and respond to wellbeing signals across an ecosystem rather than inside any single platform.
The full Architecture Note specifies the Human Web's purpose, the six entry conditions including reflection, the structure of the actors that operate within it, the patterns of organisational participation, the agent functional taxonomy, and the rules that hold the architecture together.