From one town to the Netizen Stack

Röbel was never the goal. It was the proof. A 750-year-old town now verifies citizens without collecting a single document, votes in secret on Base, and opens its digital ballot box only when three of five citizens turn their keys together. It runs on $15 a month. One person operated it while building it.

The question we get asked, and the question we ask ourselves, is what this becomes. Here is our current thinking, stated plainly so we can be held to it.

The problem, named precisely

Communities digitize through platforms that surveil by default. Towns buy citizen-engagement SaaS that knows who clicked what. Unions and cooperatives vote by email or a web form with an admin panel that sees every ballot. DAOs vote in public with tokens, which imports plutocracy and exposes every member's position forever. The one mechanism every serious community relies on, the secret ballot with a trustworthy count, has no good digital form.

The pieces to fix this have existed in the Ethereum research community for years: MACI for private verifiable voting, account abstraction for invisible onchain UX, cheap L2 blockspace, threshold cryptography for removing single operators. What did not exist was the integration, the operational playbook, and the language. That is the actual product surface. Cryptography papers do not onboard a 67-year-old council member. We now know what does.

What the Netizen Stack is

Three layers, each independently useful, together a complete civic operating system:

1. Citizenship. Sybil-resistant, document-free identity by social attestation. Soulbound credentials, community-governed thresholds, public revocation trails. Who belongs is decided by the community, recorded onchain, and surveils no one.
2. Voice. MACI-based secret ballots with proven-correct tallies, wrapped in account abstraction and deterministic keys so that voting is one tap, costs nothing, and never shows a wallet.
3. Trust. No single operator. The coordinator key lives as a Shamir federation across community members. Tallies require a quorum acting in public. Between elections the key does not exist.

The name says the thesis: modular infrastructure that communities assemble, in the way successful open stacks in this ecosystem are assembled, rather than a monolithic app they rent.

Product, platform, or protocol

Our honest assessment, in order of when:

Now: open source reference plus a hosted pilot practice. The stack stays open (it already is, AGPL). We earn by doing what we just did for Röbel: deployment, key ceremonies, operations, and the unglamorous work of words and workflows. Three to five lighthouse communities, each a public case study. This is consulting-shaped revenue and that is fine; at this stage the scarce asset is proof, not margin.

Next: governance as a service. The repeatable parts are already visible in the Röbel runbook: contract deployment, coordinator hosting, ceremony tooling, the keyholder dashboards, the citizen apps. Packaged, that is a SaaS a town clerk or a co-op board can adopt without us in the room: setup wizard, hosted coordinator with the federation owned by the community (we host machinery, never keys), white-label apps, compliance documentation. Price like civic software, anchor on cost displaced (a single postal ballot round for a mid-size Verein costs more than a year of this).

Later, maybe: a protocol. A shared attestation and ballot layer that communities join rather than deploy could be valuable, and it is the only stage where a token question is even coherent. We are explicitly not there. A token before product-market fit would be a distraction at best and a credibility tax at worst, particularly with the audience we serve: German municipalities do not buy tokens, they buy outcomes and references.

The market, soberly

Germany alone has roughly 11,000 municipalities, 600,000 registered associations, and thousands of cooperatives, unions, and housing societies. Nearly all of them vote, regularly, expensively, and on paper. The EU's digital-sovereignty agenda is pushing money toward exactly this gap, and the existing e-participation vendors offer engagement analytics, not secret ballots. On the crypto-native side, every serious DAO conversation eventually reaches private voting and stalls on the coordinator problem we just solved in production. We can serve both, with the same stack, and each audience makes us more credible with the other.

The honest risks, so nobody thinks we have not noticed: binding legal elections are years away and gated by regulation we do not control (we position as the layer for everything below that threshold, which is most community decision-making); key ceremonies need committed humans, which is a community-readiness filter as much as a feature; and our numbers are pilot-sized. The answer to the last one is not bigger claims, it is more towns.

On tokenization and AI

Both come up in every conversation, so here is our position. Tokenization is a natural adjacency: once a community has verified members and governed treasuries, community assets (a solar roof, a festival, municipal projects) can be financed and owned through the same rails. It is on the roadmap as a layer above the stack, after governance is boring and reliable. AI is already inside the product (Röbel ships an assistant citizens actually use) and the long-term shape is obvious to us: as agents act on people's behalf, proving there is one real human behind an action becomes the scarce primitive. A citizenship layer that can answer "is this a verified person of this community" without revealing anything else is infrastructure AI will need, not the other way round.

Neither of these earns the right to distract us before the core loop (verify, vote, trust, repeat) runs in ten communities without us in the room.

What we are doing at Berlin Blockchain Week

Showing the receipts. The federation ceremony, the production tally pipeline, the day our system proved an honest zero rather than fabricate a result. If you work on MACI, account abstraction, attestations, or you know a community that should run on this, find us. The stack is open. The first town is live. The next one is the point.