Zero Trackers, 27 Languages: Gewerkton’s Egress-Free Architecture Treats Privacy as Infrastructure

A marketing site translated into 27 languages would usually be an invitation to complexity. More markets can mean more analytics, consent tooling and third-party services, each adding another connection between the visitor and an outside system. Gewerkton has taken a different route: its marketing site has zero trackers, no cookie banner and a fully egress-free architecture.

That is more than a tidy privacy statement. It is an architectural choice that reflects the purpose of the product behind the site. Gewerkton is a voice-first construction documentation and defect management platform for global markets. Its job is to turn activity on construction sites into evidence, coordinate that evidence across tools and teams, and preserve the relationship between what happened, what was recorded and what was reported.

For a platform built around proof, data custody cannot be an afterthought. The privacy engineering of the public site, the choice between an EU cloud and an organisation’s own infrastructure, and the ability to select AI providers by region all belong to the same architecture story. They address different layers, but each asks the same question: who controls where information goes?

The product is in beta now, with a public beta planned for fall 2026. Its current status matters because this is not a story about a finished platform being presented as settled infrastructure. It is a look at the data stance being built into the system while the product is still taking shape.

The quiet significance of an egress-free marketing site

Gewerkton’s marketing site combines 27 content languages with zero trackers, no cookie banner and a fully egress-free architecture. Those details are easy to read as separate features. Together, they describe a deliberate boundary.

A tracker-free site does not create a stream of behavioural data about its visitors. The absence of a cookie banner follows naturally from the stated architecture rather than serving as a cosmetic gesture. The egress-free design completes the picture: the site is structured around keeping its operation free from outbound data flows.

This is notable because international reach often becomes an argument for collecting more data. A multilingual site may be expected to measure regional traffic, compare audience behaviour or assemble increasingly detailed profiles. Gewerkton demonstrates another possibility. It can present material in 27 languages without making visitor surveillance part of that presentation.

The distinction is important. Privacy engineering is strongest when it reduces the need for consent and disclosure machinery by changing the underlying system. A banner can explain or request permission for data processing. An architecture can avoid creating that processing path in the first place. Gewerkton’s site takes the architectural route: zero trackers, no banner and no egress.

That decision also keeps the public face of the company aligned with the platform’s core premise. The marketing line is direct: “On site, what counts is what’s proven.” A company making that argument has good reason to be precise about its own data flows. Evidence gains value from clarity about its origin, movement and custody. The public website applies the same discipline at a smaller scale.

Privacy across 27 languages

The 27-language footprint is not an ornamental translation exercise. Gewerkton is intended for global markets and cross-border teams working across the EU, US and APAC. Its deployment fields include projects in Asia with Chinese, Korean and Vietnamese crews, where information can remain multilingual from capture to report. Teams can work in their own language while the evidence original stays unambiguous.

That combination changes the privacy question. Language accessibility determines who can participate in a workflow, but regional differences also affect where teams may want their information processed and held. A global platform therefore needs more than a single language layer placed over a single data path.

Gewerkton’s answer is to separate global usability from a forced global destination for data. The site can communicate in 27 languages without trackers. The platform can support multilingual work while offering a choice of data residency. AI processing can be directed towards providers in a selected region. These are distinct controls, but they share a consistent principle: international reach does not have to mean surrendering control over routing and custody.

The platform was born in the German market and has its deepest German commercial integration through GAEB, REB, XRechnung and DATEV. At the same time, it is designed for global projects and includes regional AI-provider choice across the EU, US and Asia, including mainland China. The result is not a regional label stretched over an international product. It is a platform that combines a specific commercial foundation with a wider, selectable data architecture.

Identity-Native Infrastructure Access Management: Preventing Breaches by Eliminating Secrets and Adopting Zero Trust

View Latest Price

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Data residency is a custody decision

Gewerkton offers a straightforward data-residency choice: the EU cloud or the customer’s own infrastructure. The significance lies in making that choice part of the platform’s design, not in presenting one hosting model as universally correct.

Some organisations may choose the EU cloud. Others may want the system inside their own house, under their own infrastructure decisions. Gewerkton supports both positions. That gives teams a way to align the deployment with the project rather than forcing every project into the same custody arrangement.

This matters particularly for construction evidence because the material is tied to events, responsibilities and decisions. A dictated observation can become evidence. A defect can be connected to a photograph and a deadline. A daywork report can be dictated and signed on the device at handover. An instruction can remain backed by the original audio. The platform is not merely moving generic office content between screens; it is handling records intended to establish what occurred on site.

When that is the job, storage location becomes part of the product’s operating model. The question is not only whether information can be retrieved. It is also whether the organisation can choose the environment in which the information remains. An EU cloud and self-hosted infrastructure represent two answers within the same platform.

Gewerkton Cloud carries this architecture across the product line. It provides operations and model/data coordination between Field, Studio and third parties. Because it sits at that coordination point, its data stance affects more than one interface. It shapes how material from site capture, browser-based plans and models, and external participants can be coordinated without reducing every deployment to one fixed residency choice.

Data Security in the Age of AI: A Guide to Protecting Data and Reducing Risk in an AI-Driven World

View Latest Price

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Regional AI selection without a compulsory provider

AI introduces a second routing decision. Gewerkton’s BYO-AI model supports 13 AI providers, allows organisations to bring their own keys and makes the processing region selectable. Options span providers in the EU, US and Asia, including mainland China.

The important feature is not simply the number of supported providers. It is the separation of the platform from a compulsory AI destination. A team can select a provider and region instead of being locked into one vendor. Bringing its own keys further places that relationship under the organisation’s control.

This approach fits a platform used across different markets and project structures. A cross-border project may include teams in the EU, US and APAC. A project in Asia may involve Chinese, Korean and Vietnamese crews. The same software can support multilingual capture and reporting while the organisation chooses an AI provider suited to the region in which it is operating.

Regional provider choice and data residency should not be confused. One concerns where the platform is deployed and its data held; the other concerns the provider selected for AI work. Gewerkton exposes choices at both layers: EU cloud or the organisation’s own infrastructure for residency, and 13 bring-your-own-key providers across the EU, US and Asia for AI.

That separation makes the architecture legible. “Global” does not describe a single invisible processing route. It describes a product that can operate across regions while letting users make explicit choices about infrastructure and AI providers.

Evidence begins at the edge of the site

Gewerkton Field is the voice-first construction site app. It covers dictation to evidence, defects, daywork reports, takt and portal workflows. This is where the platform meets the conditions of active sites: distributed work, rotating crews, parallel trades, handovers and instructions that may need to remain connected to their original audio.

In wind farms and renewables, those conditions include distributed sites, field acceptance and offline capture in dead zones. In housing and building construction, the flow can include defects with a photo and deadline, dictated daywork reports and a signature on the device at handover. Infrastructure and tunnel projects add long durations, many change orders and instructions backed by original audio.

These scenarios show why data architecture cannot be confined to a policy page. The evidence originates in varied environments and may be captured away from dependable connectivity. It then needs to remain useful across project stages. Custody, regional processing and coordination are therefore operational matters, not abstract privacy preferences.

Voice also makes provenance particularly visible. The original spoken record and the report derived from it are related but not identical objects. Gewerkton’s stated deployment model keeps original evidence unambiguous across multilingual teams. That is central to a system in which people may capture information in their own language while collaborating across regions.

Plans, models and browser-based creation

Gewerkton Studio is the browser workspace for plans and models. Where no model exists, the site team can create one in the browser. This gives recorded site information a spatial and project context without assuming that every job begins with a complete model.

That flexibility matters across the platform’s deployment fields. Data centres and industrial plants can have many trades working in parallel under tight deadlines; meeting decisions can become trade-sorted task lists. Housing projects may organise defects around photographs and deadlines. Long-running infrastructure work may accumulate instructions and change orders. Studio provides the plan-and-model workspace, while Field handles site capture and Cloud coordinates operations and model/data flows.

The three product lines remain one branded house. Field is the voice-first site app. Studio is the browser workspace. Cloud provides the operational and data coordination layer between them and third parties. That division makes it possible to describe the platform without pretending that capture, modelling and custody are the same function.

Building the architecture with coding-agent fleets

Gewerkton is being built by a solo founder directing a fleet of coding agents using Codex and Claude. In one night, that fleet shipped 21 software packages, verified with negative controls and mutation tests.

The fact is relevant to the architecture story because it describes how a broad platform is being assembled and checked while remaining under the direction of one founder. It also places the current beta status in context. Gewerkton is in beta now, and the public beta is planned for fall 2026.

The project’s media bank follows a similarly direct production model, with more than 51 self-produced clips and posters. The public presentation may be multilingual and media-rich, but its data stance stays restrained: zero trackers, no cookie banner and a fully egress-free marketing architecture.

Privacy without theatre

There is no need to turn this into a fear story. The more useful lesson is that privacy can be expressed as a set of ordinary architectural decisions.

On the marketing site, the decision is to operate across 27 languages with zero trackers, no cookie banner and no egress. In platform deployment, it is the choice between an EU cloud and an organisation’s own infrastructure. In AI, it is the ability to bring keys, choose among 13 providers and select a region across the EU, US or Asia, including mainland China. In the product itself, it is a division of responsibilities across Field, Studio and Cloud.

Each choice reduces ambiguity about where data belongs and how it can move. None requires the platform to retreat from international work. Gewerkton can support cross-border teams, projects in Asia and sector-specific workflows while preserving explicit decisions about residency and provider region.

That is the central privacy-engineering case study. A platform built to document what can be proven treats custody as part of the evidence chain. Its public website demonstrates the principle through restraint. Its Cloud layer extends it into operations and coordination. Its BYO-AI model prevents regional processing from becoming synonymous with one compulsory provider.

Gewerkton is still in beta, with its public beta planned for fall 2026. Even at this stage, however, the architecture sets out a clear position: global language coverage, AI-assisted workflows and cross-border coordination do not require an opaque data model. For a platform whose purpose is evidence, control over where that evidence resides and which providers may process it is not an accessory. It is part of what the system is designed to coordinate.