Gewerkton: How a Solo Founder Is Building a Voice-First Platform for Construction Sites Worldwide

Construction work creates a constant stream of facts: what was completed, what went wrong, who gave an instruction, which deadline was agreed and whether a handover was accepted. Much of that information begins as speech, often while someone is standing on a scaffold, walking through a tunnel or moving between crews. Gewerkton is being built around that reality. The voice-first construction documentation and defect management platform turns dictated site information into structured evidence supported by elements such as a photo, deadline and signature.

The platform was born in the German market, where it has its deepest commercial integration through GAEB, REB, XRechnung and DATEV. Its intended reach, however, is global. It supports content in 27 languages and lets users select AI providers by region, covering the EU, the US and Asia, including mainland China. The proposition is summed up by a direct marketing line: “On site, what counts is what’s proven.”

There is an important qualification. Gewerkton is in beta now, with a public beta planned for fall 2026. It is not being presented as a finished, generally available product. What already makes the project notable is the combination of its construction focus, multilingual scope and unusual development model.

One founder directing a fleet of agents

Gewerkton is being built by a solo founder directing a fleet of coding agents based on Codex and Claude. The arrangement offers a contemporary version of the one-person company: the founder sets the direction while software agents carry out development work across multiple packages.

In one night, that fleet shipped 21 software packages. The work was verified with negative controls and mutation tests. In plain terms, those methods are designed to do more than confirm that software behaves correctly under expected conditions. They also check whether tests detect deliberately introduced faults and reject outcomes that should fail.

That technical production story matters because Gewerkton is not a single-purpose dictation tool. It is a branded house with three connected product lines: Field, Studio and Cloud. Together they cover site capture, browser-based plan and model work, and coordination between teams, data and outside systems.

Field begins with the spoken record

Gewerkton Field is the voice-first construction site app. Its functions include dictation to evidence, defects, daywork reports, takt and a portal. Takt refers to organising construction work into a planned rhythm, so different trades can move through tasks and areas in a coordinated sequence.

The practical starting point is simple: site teams can speak while they work instead of waiting to reconstruct events later. A defect can be recorded with a photo and deadline. A daywork report can be dictated. At handover, a signature can be captured on the device. The result is a structured record tied to the information that supports it.

That approach applies across very different settings. On wind farms and renewable-energy projects, sites may be distributed, crews may rotate and field acceptance may take place where connectivity is poor. Field supports offline capture in those dead zones. In housing and building construction, the focus can include defects with photographs and deadlines, dictated daywork reports and device-based signatures at handover.

For infrastructure and tunnel projects, where work can run for long periods and generate many change orders, instructions can be backed by the original audio. That preserves the source from which the structured information was created, rather than separating the report from the spoken record behind it.

XLW 195 Seconds Message Voice Recorder, Voice Recording Button Record Custom Messages for Crafts, Stuffed Bear, Baby Doll,Pillow and Xmas Halloween New Year Birthday Presents White-2P

• Recording Duration: Up to 195 seconds of audio
• Rechargeable Battery: USB rechargeable lithium battery
• Easy Operation: Press and hold to record

View Latest Price

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Studio puts plans and models in the browser

Gewerkton Studio is the browser workspace for plans and models. Crucially, its role does not depend on a model already being available. When a project has no model, the site team can create one in the browser.

This connects the immediate reality of a site with a visual workspace that can be used to organise plans and models. It also makes the platform relevant to projects that begin with different levels of digital preparation. A team with an existing model and a team without one can both enter the same product line, although by different routes.

That distinction is significant for a general construction audience. Digital building systems are often discussed as if every project begins with complete, carefully prepared information. Gewerkton’s stated design allows the people on site to create a browser-based model when that starting point is absent.

Cloud connects operations and information

Gewerkton Cloud handles operations and model or data coordination between Field, Studio and third parties. Its purpose is connective: information captured at the construction site and information organised in the browser can move within the same branded system while also being coordinated with outside participants.

For data centres and industrial plants, where many trades may work in parallel under tight deadlines, meeting decisions can become task lists sorted by trade. This turns a shared discussion into work grouped for the teams responsible for carrying it out.

The platform’s data-residency choice is similarly direct. Organisations can use an EU cloud or their own infrastructure. That choice sits alongside a broader bring-your-own-AI model rather than requiring every project to use one provider in one region.

Twenty-seven languages, with regional AI choice

Gewerkton’s BYO-AI setup covers 13 AI providers. Users bring their own keys and select a region from EU, US or Asia, including mainland China. The stated aim is to avoid vendor lock-in while allowing provider choice to reflect where a project operates and where its data needs to reside.

The 27 content languages address another construction reality: the people capturing information may not share a first language. Cross-border teams in the EU, US and APAC can work on the same project in their own languages while the original evidence remains unambiguous. On projects in Asia, Chinese, Korean and Vietnamese crews can work with multilingual capture through to the resulting report, with data residency selected by the user.

This is more than a translated interface. The supplied product facts place multilingual operation across the path from capture to report. That matters when spoken information becomes project evidence, because language support has to follow the record rather than appearing only at the point of navigation.

A deliberately contained digital footprint

The public-facing side of the project reflects the same international scope. The Gewerkton website is available in 27 languages. It uses zero trackers, has no cookie banner and runs on a fully egress-free architecture. Its media bank contains more than 51 self-produced clips and posters.

Those details give the brand a substantial body of material while keeping the site’s tracking position straightforward. They also mirror the product’s emphasis on regional choice and controlled information flows, without changing the central story: Gewerkton is intended to help site teams turn spoken events into organised records.

A broad ambition, still in beta

Across Field, Studio and Cloud, the platform follows construction information from the moment it is spoken through to plans, models, reports, tasks and coordination. Its deployment fields range from housing and renewables to data centres, industrial plants, infrastructure, tunnels and cross-border projects.

The scale of that ambition should be read alongside the product’s present status. Gewerkton is in beta now, and its public beta is planned for fall 2026. The platform is therefore still at a defined pre-release stage.

Even so, its underlying story is already clear: one founder is using a fleet of coding agents to build for a physical industry in which evidence often begins with a person speaking amid the work. With 27 languages, regional AI-provider choice and three connected product lines, Gewerkton is attempting to make that spoken record usable from the scaffold to the browser workspace and the wider project operation.