Regiothek

Regiothek builds digital infrastructure for regional food networks

Regiothek is a software company in Passau, in the southeast corner of Bavaria. We started in 2018. More than 3,000 food businesses across Germany, Austria and Switzerland now use the platform to show what they make and who they buy from and sell to (August 2026). In three regions we also run a B2B marketplace, where a restaurant or a school kitchen can order from a dozen local producers in one basket and get one delivery.

If you want to build one where you are, get in touch. We have done it three times and we are not precious about sharing how.

None of this is a demand problem. The German agriculture ministry's 2024 Ernährungsreport found that 77 percent of people look for regional origin when they shop, and that roughly two thirds expect regional to mean something like 100 kilometers. The producers exist too. What is missing is a practical way to trade between them.

That gap comes in different sizes. Sometimes it is only visibility, knowing who is out there at all. Sometimes it runs the whole way through: ordering, delivery routes, invoicing, and the coordination of an entire supply chain from the field to a professional kitchen.

Where regional sourcing actually breaks

Three situations, all of them ordinary.

The kitchen that has half an hour a week

Ask a commercial kitchen how much time it has for purchasing and the answer is often about half an hour a week. Not half an hour for the regional part. Half an hour for all of it. Inside that budget one wholesaler with one catalogue, one delivery and one invoice is not a compromise, it is the only thing that fits. Finding the farms, calling each one, tracking who has what this week, taking deliveries on four separate mornings and reconciling four invoices at the end of the month does not fit, and no amount of goodwill makes it fit. So the kitchen orders from the wholesaler, and the grower twenty minutes up the road never gets asked.

The farm with the right products and no way to reach a buyer

Three tons of potatoes is one sale. Two tons of potatoes, 3,000 heads of lettuce, 200 kilos of tomatoes and a few crates of lamb's lettuce is roughly the same tonnage and several times the work: prices to set for each line, availability to confirm every week, more things that can go wrong when the weather turns. Diversified farms are what a resilient food system is supposed to look like. They also carry the heaviest marketing burden, and mostly they carry it alone.

The region that cannot see its own food economy

A rural district knows it has farms and bakeries and a couple of dairies. It usually cannot tell you who supplies whom, what volumes are genuinely available in October, or which kitchen would buy more if somebody asked. You can fund a project without that picture. You cannot steer one.

Behind all three sits the same arithmetic. Germany has more than 250,000 farms, and four retail groups control roughly 85 percent of German food retail. Wholesale runs on the same logic: standard articles, large volumes, year-round availability. Those systems are very good at what they were built for. They move a pallet of one thing well and forty crates of eleven things badly. So the regional market stays small, and people conclude it is small because nobody wants it. We think it stays small because nobody has organized it.

What we built

Profiles and the map

Any food business in Germany, Austria or Switzerland can put itself on Regiothek. It takes about three minutes: what you produce, where you are, and, the part that actually matters, who you buy from and who you sell to. Those links are public. One profile is a listing. Three thousand profiles with their supply relationships filled in is a working map of who trades with whom, and that map is what regions, associations and journalists keep asking us for. Supply chain transparency is where the company started, and it is still how most businesses arrive.

The B2B marketplace

In marketplace regions a buyer opens one basket and fills it from several farms in a few clicks. Minimum order values are deliberately low. Everything arrives in one bundled delivery, and the month ends with one invoice to pay. That is the entire design goal: regional buying that fits inside the half hour, alongside everything else the kitchen has to order that week. The buyers are restaurants, retailers, farm shops and institutional catering, which means school and hospital kitchens and company canteens. Producers get a commercial sales channel without having to run any logistics of their own.

The delivery day is fixed. The route is not. Every week a different set of buyers orders from a different set of farms, so the orders determine the stops and the stops determine the route. Nobody drives out to a farm with nothing to hand over. Logistics is the hardest cost to control in a small-scale food system, and coordinating it properly is the only thing that makes it work at this size. That saves hours and diesel, and both of those end up in the price.

In Passau the rhythm is simple. Order by Tuesday midday, delivery Wednesday. Some of this is less convenient than a wholesaler and we say so before anyone signs up. The delivery day is fixed rather than chosen. If nobody has lettuce this week then there is no lettuce this week, which is what real seasonality looks like once it reaches an order form.

What Regiothek is not

We are not a wholesaler and we are not a food hub. A food hub buys goods, stores them and sells them on at a margin. We do none of that. There is no warehouse, we do not own a truck, and at no point do the goods belong to us. What we actually provide is the coordination: the software, the ordering logic, the weekly route planning, the monthly accounting, and the local logistics partner who does the driving.

Small volumes, wide variety, regional quirks and thin margins make this business unattractive to buy into. Organizing it is a different proposition entirely, and that difference is most of the reason the model works at all.

Nobody left this market alone by mistake

It would be flattering to read all of this as something the established players overlooked. They did not overlook it. For a wholesaler with a working depot and a full truck, taking on eighty small producers with changing availability and modest order values is a poor investment, and putting the same money into the business they already run is a better one. That is not a failure of imagination. It is arithmetic, and it comes out the same way every year, which is why nothing moves.

Which means the trade we organize is mostly not trade that moved across from somewhere else. It is trade that was not happening. A kitchen that has never bought from the dairy eight kilometers away is not switching suppliers when it places its first order with us. Something is being sold that was not being sold.

That shapes how it starts. A first delivery day in a new region is small, the volumes look unserious next to a wholesaler's, and the assortment has obvious holes in it. That is what a market looks like at the beginning. It is also why it can be built at all, because it starts underneath the level anyone established considers worth serving. Whether it stays down there is the interesting question. In Cham it has been growing for three years.

Three regions, and what we have learned in them

Across all of it: more than 3,000 businesses on the platform, roughly a hundred new ones a month, about seventy trading through the marketplaces (August 2026).

The number we watch is not the first order. Anyone will try something once. It is the fourth order, or the fortieth, because by then the products are being found, the quantities are right, and the delivery and the invoice have quietly fitted themselves into somebody's week.

We do not replace the systems that are already there

A farm may already run an ERP. A distributor may already have route planning software it likes. A kitchen may already know every producer within thirty kilometers and simply want the ordering to stop being a phone call. None of those situations calls for a new market built from scratch. They call for the pieces that already exist to be connected, and for the manual work in between to disappear.

Regiothek is built for that too. Existing supplier relationships can be organized digitally without anybody changing who they buy from. Interfaces move product data, orders, delivery information and accounting data between systems instead of somebody typing the same figures in a second time. The platform exchanges data with other systems over an API on a daily schedule, without anyone having to remember to trigger it.

There is a second effect, and regions usually notice it later. Once products, orders and deliveries run through one common structure, you can finally see the market: what is genuinely on offer, what buyers keep asking for, where the assortment has a hole, and how trade develops from one quarter to the next. That is useful on an ordinary Tuesday morning. It is also the first time a region can develop its food economy from evidence instead of impressions.

What people actually come to us with

Almost nobody arrives asking for digital infrastructure. These are the requests that reach us, more or less in these words:

Some of those are a marketplace. Some are only visibility. Several start as one and turn into the other about a year later, once people can see who is actually out there.

Regiothek can start at any of those points. No region has to work through a fixed sequence, and none of them has to want a marketplace on day one.

The market does not scale. The way to build one does.

Every region has different farms, different products and a different set of reasons why the last attempt stalled. The work underneath is nearly identical. Getting assortments into the system and keeping them current. Persuading a buyer to place a first order, which is consistently the hard part.

We solve those things once, in the platform, instead of once per region. What Cham learned is sitting in Passau. What Passau is learning now goes to whoever is next. That is why a new region can start in months rather than years, and why the price stays inside what a rural district can approve without a special budget line.

We are not trying to make one regional market as large as possible. We are trying to make the next twenty easier to build.

We would much rather do this with somebody local

In every region we work with a partner on the ground, and the division of labour has been the same everywhere so far. We bring the software, the data and the routines we have already made our mistakes with. They bring the phone numbers. The partner is usually a district administration, a regional development office, a food policy council, or a cooperative like the Maschinenring.

We can carry more of it ourselves, and at the start we sometimes do. It is just slower and it works less well. A district office gets an answer in two phone calls that would cost us ten and a drive across the county. Farmers and bakers decide whether to join based on who is asking. Anyone who tells you the software does that part has not tried it.

Where this came from

In 2016 Alexander Treml was studying computer science in Passau and selling bread on the side for a small organic bakery. Farm shop during the week, market stalls on Saturday. Wood-fired loaves, flour from a mill up the road, the same customers every week. Nobody there used the word regional. It was simply how the bakery worked, and it was invisible to anyone standing more than one step away from it.

The first prototypes were built at the Chair of Data Science at the University of Passau. Regiothek GmbH followed in 2018. In 2019 the company was recognized at the Bayerischer Gründerpreis and at the European Innovation in Politics Award in the ecology category. Then came 2020, and a large share of our customer base was restaurants. The years after that were the ones that decided what the company would actually be.

Structured data has been underneath all of it from the beginning. A food ontology from 2017 has grown into the knowledge graph the platform still runs on, linking products, businesses and supply relationships. In 2021 we worked with the University of Passau on conversational interfaces in a research project called SMAEG Bot, a year before anyone outside the field had heard of ChatGPT. We mention it because it is the reason the data now sits in one structure rather than scattered across everybody's laptops.

The company is run from Passau by Alexander Treml and Bastian Kühnel, both co-founders and managing directors, with Anton Kohlbauer, also a co-founder, on software development. The team is small and mostly computer scientists. Everybody talks to customers.

If you want one of these in your region, talk to us

That is the direct version. If a district, a city, an association or a regional initiative wants local food to be genuinely orderable in its own area, we can build it, and building the next one is what we are here for.

What you get from us is operating experience rather than a concept. The platform has been running since 2018 and the first B2B marketplace since 2023, which means the answers are specific: what a first delivery day realistically looks like, how many buyers a delivery day needs before it pays for itself, and which parts of the process will start annoying people in month four.

You also do not start from an empty map. More than 3,000 food businesses across Germany, Austria and Switzerland are already registered on Regiothek, and a fair number of them are in your area already. In most regions we have looked at so far, the first usable list of candidate producers came out of the platform the same afternoon somebody asked for it.

And the software is ours. Regiothek is built by computer scientists in Passau rather than resold through a partner network, so when a region needs something the platform cannot do yet, we can go and build it. Cham needed a different collection logic than Passau. Nordwest needed something else again. That work happens in house, and then it belongs to every region on the platform.

One thing we say early, because it saves everyone time. A regional B2B marketplace needs a few years to carry itself, not a few months. A project that runs twelve months with nothing planned behind it does not get there, and the region walks away concluding that regional trade does not work, when what actually happened is that the funding ran out.

Half an hour on a call is usually enough to tell whether it would carry where you are. Write to contact@regiothek.org.

The rest of it is in German

We write in German, and this page is the short version of several years of it. If you read German, these go further:

Regiothek GmbH, Passau, Germany. regiothek.de · Imprint