The federation uses technology so that human energy can go where only humans can go. What software handles, hands and minds are freed for. What the machine cannot do, presence, judgement, relationship, honesty with oneself, stays human work.
A federation of sixteen members across one , with a public site in seven languages, a registry of decisions, a cooperative legal structure, and an intention to scale through other Labs, cannot be carried by hand alone. Documentation drifts. Translations fall behind. Decisions get lost in chat threads. Without infrastructure, a community of this kind spends most of its energy on remembering what it already knows.
What technology does, when it is correctly chosen, is take that load. It remembers. It translates. It registers. It makes what one person did visible to the next. It makes what was decided last spring still findable in autumn.
Living together is human work. Making decisions is human work. Carrying conflict, repairing trust, deciding what to plant where, holding a fire-watch in summer, tending bodies and meals and children, sitting with a member who is leaving the Lab. The federation uses technology to free human energy for that work. To make space for what is irreducibly human, not to replace it.
We use technology so that what is uniquely human can be lived more fully. Not less, more.
Three layers stand: the federation site, four instruments where the work becomes legible, and a document pipeline. Each is open source. Each is checkable.
The federation site (syntrociety.org)
The site is the public face of the federation. It hosts the , the Practice, the , the Quadruple Helix description, every essay, every Lab page. It runs in seven languages with English as the source.
The site is built with Next.js 16 (App Router) on React 18, written in TypeScript, styled with Tailwind CSS, and hosted on Vercel. Most pages are pre-rendered at build time and served as static HTML; a small number of routes (API endpoints, dynamic OG images) run on demand as serverless functions. The repository is open and can be mirrored, archived, or self-hosted by any party that wants to verify what it says. There is no proprietary database that holds federation content; everything that publishes is in the source repository, which is itself open. A reader who wants to know what changed and when can read the commit history.
What this means in practice: the federation's words live in a place where they can be checked, copied, and continued by others. If syntrociety.org disappeared tomorrow, the content would still exist in repositories, archives, and mirrors. The federation does not own its own infrastructure as private asset; it stewards it as common ground.
Four instruments · people, chart, compass, lens
The federation builds and runs four instruments. Each is open source. Each is checkable. Together they hold what the federation does over time.
people · the identity host (people.syntrociety.org)
people is the federation's member registry, and the identity layer for everything else. Whoever logs in at people is the same person at chart, compass, and lens. Members, Friends, seats, and consents live here. If a Friend withdraws consent, it happens in one place. If a Lab removes a Member, it happens in one place. No fragmented accounts.
chart · the records platform (chart.syfers.eu)
chart is where proposals, decisions, tasks, meetings, plantings, energy, reflections, and events are kept. It is multi-tenant: each Lab has its own subdomain and its own records. Council decisions are written here with their attendees, objections, and amendments. After consent, a proposal is locked; subsequent revisions enter as amendments linked back to the original.
compass · the self-assessment (compass.syfers.eu)
compass lets a Lab place itself, for each of the sixteen conditions, on one of four levels: absent, emerging, established, embedded. A mirror, not a gate. Nobody scores anyone. The Lab tells itself the truth about where it is, in the same words every other Lab uses.
lens receives documents by email or upload: letters, contracts, reports. It classifies, translates where needed, and archives them in a legible, searchable form. A letter from a municipality in Portuguese can be read by a Council member in English the same afternoon. The original and the translation are sibling documents in the archive, linked by canonical id.
How the four work together
The four instruments do not produce the federation. They are where what the federation already does becomes legible: to the Lab itself, to other Labs, and to whoever reads it twenty years from now. Without them, the work happens and disappears.
A letter arrives in lens. A Council reads the translation, takes a decision, and the decision lands in chart. The Council members did not log in three times. Their account in people is the same account that signed the decision in chart and read the translation in lens. Two years later, the Lab looks at itself on the condition inclusive governance in compass. The self-assessment is not an interpretation after the fact. It is the documents in lens and the proposals in chart, read back.
One identity. One archive. Four pillars.
Document generation
Most federation documents (this page, the essays, the Charter revisions, the Lens, the briefings with Claude Code) are produced through a working pattern that combines human authorship with AI assistance. The pattern is described in § 03. The output is markdown source, version-controlled, rendered to the public site, and exportable to PDF for circulation.
The federation does not hide that this pattern uses AI. It also does not credit AI as author. The judgement of what is true, what is honest, and what serves the work belongs to the people whose names stand on the documents. The pattern itself is described openly so that other Labs and other initiatives can replicate, adapt, or critique it.
AI assistance plays three roles in the federation's work. Each is bounded. Each is checkable. Each leaves the human judgement intact.
Writing
The first role is conversational. AI is used as a thinking partner for the formulation of texts: federation essays, briefings, this page, charter revisions, translations between English, Dutch, and Portuguese. The pattern is dialogical. A human brings the question or the draft; the AI offers formulation, alternatives, and reflection; the human chooses what to keep. AI does not author federation documents. It helps articulate what the people in the federation already know but have not yet put into words.
This role is the most visible and the least unusual. Writers, researchers, and policy-makers across Europe have been using AI in this way since 2023. The federation makes the use explicit so that what is read can be verified.
Building
The second role is constructive. AI assists in the production of code, of site structure, of document generation, of translations, of design assets. This page, for example, was drafted in conversation with Claude. The site you are reading was built with Claude Code, an AI-assisted development environment, by a human developer who specified what was needed, reviewed each commit, and approved each change.
Here too, the human stays the agent. The AI suggests; the human accepts or rejects; the change is committed; the change is reviewed; the change goes live. Every line of code on syntrociety.org has passed through this loop. There is no AI-generated content that has not been read by a person who understood what was being said and chose to publish it.
Reflecting
The third role is the one that makes the federation's use of AI distinct. AI reads the data the federation has registered in SYFERS, and it asks the federation what its own data shows. Not from a neutral position. From a specific perspective: the patterns that living systems show. These are documented in The Lens, the editorial reference at /lens. For the working analysis in SYFERS, the federation uses Patterns, an operational subset of Lens entries with weights set per Council or per analysis.
The pattern works as follows. SYFERS holds the federation's record: decisions, contributions, member activity, project progress, financial flows. The Lens describes the patterns that living systems show: circulation versus accumulation, reciprocity versus extraction, distribution versus concentration. Patterns is the operational subset of these in SYFERS, with weights per analysis. AI is given the task of reading the data through Patterns: where does federation behaviour show circulation, and where does it show accumulation? Where does it show reciprocity, and where extraction?
The output is not a verdict. The output is a question, returned to the community for its own quarterly reflection.
An illustrative example of such a question, as it might come back to the community on a quarterly basis:
In the past quarter, decisions with financial implications were prepared by the same three members in 80 percent of cases. The Lens entry on Mandate describes rotation as a condition for healthy distribution of authority. What does the community make of this pattern?
That question goes to the group. The group decides what to do with it, or whether to do anything.
This role is the federation's instrument for what Charter-principle Truth asks: to be honest with itself. A community of sixteen could in principle do this work by hand. In practice, no one has the time. AI makes feasible what otherwise stays an aspiration.
The reflecting role asks the most careful design. The federation has built four safeguards into the pattern. Each is operational, not aspirational. Each is checkable.
Safeguard 01
Pseudonymisation with traceability
Data that becomes public in any analysis is pseudonymised. Members are not named in outputs that circulate. A mapping between pseudonyms and identities exists, is maintained, and is not public. This allows a community member to trace a finding back to its source if needed; it prevents external actors from reading personal patterns from federation data.
This is a deliberate choice over full anonymisation. Anonymisation would destroy the mapping and make traceability impossible. Pseudonymisation keeps the chain intact while keeping it private. It is GDPR-conform. It is also methodologically grounded: the federation can investigate its own findings without exposing its members.
Safeguard 02
Personal versus collective reflection
A member who asks AI for personal reflection on their own contribution data receives a personal response. That response is not shared with the community. The member chooses whether to bring something from it into a group conversation.
This applies the principle of consent to data about oneself. The federation cannot read into individual members' reflections. The individual chooses what becomes collective. The pattern protects against a subtle form of pressure that could otherwise emerge: the assumption that reflection-results are shared by default.
Safeguard 03
Quarterly reflection without blame
What goes to the group is the collective view: patterns visible in the federation's overall data, read through the Lens. The format is structured to make pattern-recognition possible without question of fault. What is visible in the data? not who caused this?
This matters because the same data can be read in two different registers. In Q1, financial decisions were concentrated among three members can be heard as accusation, or as observation. The format chooses observation. What the group then does with the observation is its own question; whether to redistribute, whether to reflect on why, whether to leave it because it served. The instrument does not push toward action; it makes pattern visible.
Safeguard 04
Patterns-guided analysis, not generic analysis
AI does not analyse the federation's data from a neutral position. There is no neutral position. AI reads the data through Patterns, the operational subset of The Lens with weights per analysis. The prompt that instructs AI is itself a federation document, available for inspection in the SYFERS code repository. A reader who distrusts the analysis can read both The Lens (the editorial source) and Patterns (the operational set) and the prompt itself, understand what AI was asked to look for, and decide whether the framing is sound.
This is methodological transparency at a level rare in algorithmic systems. Most AI-assisted analysis does not disclose its prompt. The federation's pattern asks for it: if Patterns is the interpretive frame, the frame should be public.
The instrument supports awareness; the work of being honest stays with the community.
The four safeguards make the reflecting role workable. What they do not make automatic is the work that follows reflection. That work is human.
When AI returns a question to the community, the community decides whether to take it seriously. The data shows what it shows; honesty about what to do with what it shows is a choice that no instrument can make for the people who must live with the consequences. A community can have a working AI-spiegel and choose, every quarter, not to be honest with itself about what it sees. The instrument cannot prevent that.
What the relational work asks, AI cannot do. Sitting with a member whose contribution has dropped because something difficult is happening at home. Holding a circle in which two people who have hurt each other speak to each other. Deciding that a Lab member will leave, and helping that be done with dignity. Carrying news of a death. Welcoming a new member into a kitchen in winter. None of this is in the data; none of this can be carried by an instrument.
The federation uses technology for the parts of its work that benefit from technology. Where presence is the point, presence is the practice.
Three criteria guide what the federation builds and adopts.
Tools must be checkable. Source readable, behaviour reproducible, decisions traceable. A tool whose workings cannot be examined cannot be improved, cannot be trusted in disagreement, and cannot be inherited by future Labs that want to understand what they are using. This is not a moral preference; it is a methodological one.
Tools must keep the federation's options open. Open source first; open standards always; vendor dependency resisted where it can be resisted. Where commercial services are used because they perform the work, the federation keeps exit paths open. The site can be self-hosted. The data can be exported. The pattern can be reproduced. Decisions made today should not foreclose decisions made by the federation a decade from now.
Tools must respect what members have not consented to. No tracking that members did not register themselves. No automated behavioural analysis that runs without their knowledge. The four safeguards described in § 04 are how this criterion is operationalised in the federation's own AI-assisted reflection. Tools the federation adopts from outside meet the same standard, or they are adapted before use.
What individual Labs choose for their own communication, presence, or workflow is each Lab's decision, guided by Charter principles and not by federation prescription. The federation does not legislate platform choices; it asks that whatever a Lab uses, it uses with the same discipline of openness and care that applies to everything else in its work.
The infrastructure that runs the federation's site, registry, and AI-assistance has a footprint. Servers consume electricity. Data centres consume water. Training a large language model consumes both at scale that is hard to defend on its own.
The federation's response is not to claim that its choices are clean. They are not. The choice is between footprints, not between footprint and no-footprint. Running a federation by hand, without technology, would consume different forms of energy: human energy, time, travel, paper, repeated meetings to remember what was already decided. The choice is which footprint serves the work better.
Sulitânia, the founding Lab, generates 22.3 kWp of solar electricity locally. Not enough to power its own data centre, were it to run one. The federation's infrastructure is currently hosted on commercial providers whose energy mix is not under federation control. This is a known compromise.
What the federation does choose, where it can:
Static site generation over server-side rendering. Less compute per page-load.
Open source models where they perform; commercial models only where the work needs them.
Smaller models over larger when the task allows.
Self-hosted where capacity exists; commercial hosting where it does not.
What the federation cannot solve, the federation names. To use AI is to make use of infrastructure built with significant resource cost. To not use AI is to consume different resources, often in ways that do not show on a balance sheet. The federation chooses, transparently, to use what serves the work, and to keep watching whether the choice still holds.
This is not unique to technology. Federation work is full of choices between imperfect options. To name a trade-off rather than pretend it does not exist is the first step in handling it well. Perspectives meet here, rather than being chosen between.
The pattern described above does not exist in isolation. It is part of how the federation does its work in five specific ways.
The Lens describes; Patterns weights; SYFERS records; AI reads between them. The four are one feedback loop. The Lens describes what living systems show. Patterns is the operational subset with weights per analysis. SYFERS holds what the federation has done. AI compares them and asks the question that the comparison generates. Charter-principle Truth makes the question worth asking. Practice Discipline I (Open registration) makes the data available. Practice Discipline II (Conflict is welcome) makes the answer something the community can carry.
The technology is federation-infrastructure, not Lab-infrastructure. What stands here is built for the federation. Other Labs that join will use the same SYFERS, the same Lens-prompt, the same documentation pipeline. The cost of the infrastructure does not fall on each new Lab; the federation carries it. The benefit of the infrastructure compounds with each new Lab.
The work is open and replicable. Source code is public. Documentation is public. The Lens-prompt is public. Other initiatives, regenerative communities, cooperatives, or municipalities, can read what the federation has built and adapt it for their own work. The federation is not protective of its tools. The work was made because it was needed, not because it was profitable.
The model is open for evolution. The technology described here is a first proposal. The federation invites refinement from the four helix actors: researchers who can deepen the methodology, civil society initiatives who can stress-test it, public authorities who can verify GDPR and AI-Act conformity, economic actors who can assess practical viability. The Charter and the federation's responsibility to its own members and to individuals remain leading: instruments support awareness, not control.
The Lens, Patterns and the prompt that instructs AI are all federation documents. Like Charter and Practice, they are something the federation owns, revises, and discusses. The prompt will be published in the SYFERS code repository LINK TBD.