I hope REWE is seeing this and offers an official MCP server. In the end we pay real money there. :D
I wrote a skill some time ago to support me with "agentic groceries" on my own - it's the future of shopping I would say.
My workflow:
- I paste in urls or text for receipts I will cook this week
- agent extracts the ingredients, calculates cups to ml and such, replaces meat with vegi ingredients, replaces some other things I prefer always (often also creates a nice markdown receipe at this steo I can put into Obsidian)
- check my list of favs, check search cache (so not every time the api is called, I'm a good netizen :D )
- ask me which items I have at home (no need to add to the basket)
- search rewe api for multiple candidates and let me choose.
- after each recipe I enter /new to start with fresh context
- also I have a list of things I buy every week
I still put everything manually in the basket in the end, but this is not the thing which is time intensive.
This is also my exact workflow here. Thank you. I'm adding things that should be in the basket via Siri to my Obsidian notes. I also add recipes to that list or anything else. Then Claude checks that list once I want to do the shopping, looks through my favorites, and fills the baskets with all the items needed. I also use Claude to list the most frequently bought items for a template that always gets filled in.
It is extremely simple. I just tell Claude: "Use korb to look at my order history and create a table with my most frequently bought items. Next time, fill my basket with the items I always buy."
So something of that variation works.
I do have a suggestion for your app though:
Have it compare your basket of goods across different markets in your region to show you the cheapest option.
I'm pretty sure this possibility is actually one of the reasons they locked down the API.
I've used Data from REWE in the past and made a comparison between a couple of cities in Germany (I believe it was Frankfurt, cologne, Berlin, Munich and Hamburg). Hamburg was by far the most expensive, often as much as 10-20% more expensive.
> I do have a suggestion for your app though: Have it compare your basket of goods across different markets in your region to show you the cheapest option.
This is a great idea. I just think the use case is not that big since REWE is the worst in the price/quality ration and just going to another shop would save you more.
>I do have a suggestion for your app though: Have it compare your basket of goods across different markets in your region to show you the cheapest option.
I'd settle for just being able to sort items by unit price... I'm sure this is a [regulation-]solved problem in Germany though
Sorry, yes, I'm not German and haven't used it. It was an idle complaint about how trying to use grocery store (or other similar) sites is difficult because they prevent you from being able to sort properly, for example by unit price. Sometimes they change the displayed unit per product so you also have to convert them to compare manually (less of a problem for metric, but like, drink brand 1 €2/100ml, drink brand 2 €13/1L etc).
As I was writing it, I realized that this kind of tactic just feels like it would be banned in Germany
Stores are required by law to provide the price per unit/weight/volume alongside the price, so you can directly compare the price of a pint of beer to the 0.33 liter bottle without calculating anything.
I just checked and REWE only lets you sort by absolute price. But honestly, you can compare prices so much better on their website than in a physical supermarket already [0].
What I suspect though:
They mainly show current discounts. The REWE API exposes those as a separate list for each market.
There's around 3.5k markets and each can set their own discounts and has their own product catalogue with their own pricing.
So it would be 3.5k API calls to fetch all offers for all markets. Which is doable.
But fetching all products takes like 100 calls per market. It's quite a bit of data. And I think most supermarket don't publish their catalogue at all since they don't have delivery options.
> The suggestion engine (korb suggestion threshold) is re-implemented in Lean 4 with five mathematically proven properties: suggestions have positive frequency, are sorted descending, come from ordered and available products, exclude basket items, and respect the count limit.
This is amazing. I mean this literally in that I am amazed, but also figuratively in that I am delighted to watch the type-safe-mad build cool shit.
As a SWE at Rewe (at a completely different department), I can say that I find this pretty cool. I wonder if this is going to be a wakeup to management to relax the API restrictions.
That's a good idea and actually was my first thought too. I sent a chat message so the product owner in charge is aware of the interest in something like it.
I think a big issue here is the lack of standard - there is no established way of where an MCP server should be hosted so that agents are easily able to find it. Right now, the best solution I could think of would be to serve it at something like rewe.de/shop/mcp and you'd manually have to register it with your agent.
Nice thank you. If something came out of this that would be amazing. I am not the biggest fan of mcp, as they are a waste of tokens though. I shot you a message on LinkedIn as well, was wondering if this caused some discussion internally at Rewe because it was immediately blocked.
Even a CLI interface would be better than the sorry excuse of Asda's website. I wonder if entrusting an LLM is worth the trade off with the tedium of online shopping.
I want to add something else to this. In the process of writing this, I also played with formal verification and formally verified the suggestion engine, which was a really nice side discovery.
The basic idea is to write a prove in Lean4 and then test both the production implementation (Haskell) and the Lean implementation against random inputs.
Compare if the results are the same.
If that is the case -> you can be pretty sure the unproven production version is as correct as the proven lean version.
Surprised how little the B2C and even B2B e-commerce segment is providing API access for automation and agentic coding. One could easily set up rate limits, fraud detection and KYC checks upfront initial access.
Normal customers are not paying for APIs, so there is little value in offering them for free, especially when it can be harmful. B2B sometimes does pay for them, but not always is it obvious from the outside that they even exist. Much commerce is also on tight numbers and thus kinda secretive about their stuff.
Think it's context dependent whether it's a good or bad thing.
The owners of German supermarket and car companies are really the richest of the rich in Germany (okay and maybe the SAP guy on top). It would definitely be a net positive if someone manages to scrape and compare their prices.
In the restaurant market it's one player abusing many small players.
And honestly, I think the reason everyone cries when "Amazon launches an API" is because Amazon would not dare to piss off the German supermarket oligopoly.
Can you share them? I recently looked for such projects and didn't really find anything that works well.
The issue is that each market sets their own prices and I believe REWE is the only large one where you can fairly easily scrape the product catalogue. I thought about it in a shopping list context, so you'd need to make it location dependent to be useful. But you could do a lot of cool things with it. Like choose a basket of goods and it creates a route for you: "Go to supermarket A and buy goods XYZ, then to supermarket B and buy ABC"
There's this one which got some publicity, doesn't seem to be updated any more but it worked for all these retailers listed and is open source: https://github.com/badlogic/heissepreise
> B2C: Is it really surprising that a busines has no interest in providing more price transparency to their customers?
Might I suggest you remove your tin-foil hat and consider that:
- 99% of REWE customers almost certainly have no clue what an API is
- 99% of the remaining 1% know what an API is, but their day-job involves messing with APIs, so they don't want to spend their weekend-time messing with the REWE API, they just want to do their shopping at REWE.
- The final 0.1% are those who come on HN and pretend its all some sort of big conspiracy to minimise transparency by $evil_corp. :)
If you think about it, imagine if REWE officially exposed an API B2C. This would mean they are obligated to provide support.
Do you really want the price of your shopping to increase because REWE now needs to find money to pay for a helpdesk for the millions of B2C API users ?
Businesses and services differentiating between B2C and B2B is nothing new, that is why the two different terminologies exist !
What next, you don't want to fill up your car at the petrol station (B2C) but you want to be permitted to buy a barrel crude oil direct from the drill and refine it yourself (B2B) ?
> Might I suggest you remove your tin-foil hat and consider that:
First up: Read and follow the rules. No need to insult me. Especially considering what you said shows that you both misunderstood AND misrepresented what I've said.
And frankly, my reasoning was simply saying "Company won't publicize internal info if they don't get an advantage from doing so". It's literally the same reason Google doesn't publish all of their source code. I'm struggling to see what part you are misunderstanding but it has to be something extremely basic to conclude I'm a conspiracy nut for basically stating "Company acts in their interest".
Opening an API to the public allows third parties to develop apps that can then be consumed by end-consumers. Not trying to be offensive here, but do you know what an API is? To conclude I meant every single end-consumers building their own app is at best disingenuously twisting my words.
Opening the API would allow new players like you and me to enter the market and take a piece of the pie. Why would a market, dominated and controlled by a few big players, opt for that? You don't even need to know that the German grocery market is incredibly price competitive, to understand that.
> If you think about it, imagine if REWE officially exposed an API B2C. This would mean they are obligated to provide support.
Can you provide a source for that requirement? I'm pretty sure you just made that up.
> Businesses and services differentiating between B2C and B2B is nothing new, that is why the two different terminologies exist !
At this point I'm entirely lost what you read in my comment. Yes I know. I specifically made that distinction.
> What next, you don't want to fill up your car at the petrol station (B2C) but you want to be permitted to buy a barrel crude oil direct from the drill and refine it yourself (B2B) ?
Yeah you definitely misunderstood something... What I said/meant:
The question: Why isn't the API open?
My answer:
For B2B I gave an example where the API is used by another German firm, providing an example that the API is indeed consumed B2B.
For B2C: They have no reason to do so. They have a well functioning app where you can order stuff. They have one of the bigger recipe pages (at least it does very well SEO-wise) in Germany where you can immediately order ingredients from a recipe. The biggest recipe page in Germany (chefkoch) offers a direct link from recipes to their order page. Maybe you're missing this info? Thinking it's an internal API to data that isn't exposed anywhere at all would somehow explain whatever you tried to say here. But again, if you're that uninformed, don't insult people.
> Opening an API to the public allows third parties to develop apps that can then be consumed by end-consumers. Not trying to be offensive here, but do you know what an API is? To conclude I meant every single end-consumers building their own app is at best disingenuously twisting my words.
Here you are wrong too.
If you want to develop an app via an API that is only offered B2B, what do you do ?
Yes, that's right ...
You phone up REWE and negotiate a license to access to the B2B API to develop your application. B2B2C if you want to put it in simpler terms.
My original point stands. REWE clearly do not want to officially expose the API B2C, almost certainly for the exact reasons I already spelled out in my original post.
But no, its easier for you just to spread FUD, claiming "busines has no interest in providing more price transparency to their customers" just because they will not let you have access to the API as direct B2C.
Haha man, I think this is a cool project, the REWE API is cool, the REWE delivery App and Website are cool. Certainly not spreading fear, uncertainty or doubt.
What you describe as B2B2C is exactly what chefkoch does. And it's exactly what I initially said, so I'm not sure what point you're trying to make. But anyway. Doesn't feel like we're getting anywhere. Have a great day ;)
They will lose. The only areas Germany can compete, are that with broad APIs. Take HBCI/FinTS for banking. Go API or go out of business. I‘m pretty sure Piknik and Flaschenpost will offer agentic agent compatible APIs at some point.
I mean, fixing small issues is not a big deal – during my ordering sessions, if something comes up, I actually just let Claude create an issue for it, and then when I have time, I create a fix.
I love the idea of a CLI for groceries. Do you have plans to support 're-order' scripts or meal-plan integration? I can imagine a workflow where a recipes.yaml file gets piped into your CLI to automatically fill the cart with everything needed for the week. Much faster than clicking through a mobile UI.
I remember a friend and I in college were looking into ways to do this in the US but major grocery chains here are pretty sensitive about their product data being accessible by open APIs and web scraping...
Kroger does have an open API with some limitations that get you pretty far. Someone built a python wrapper and MCP for it awhile back. [1]
The workflow in our house is basically to use Siri to add items to an iOS Reminders list as the week goes on. Then, the day of or day before we plan to shop we fill-in with commonly purchased items or one-off things and go shop.
It's been on my TODO list to have an agent "skill" take the Reminders list, ask for additional items, and populate the cart to schedule a pickup. We tend to prefer going in the store though to browse the produce because our local Kroger isn't great about that. Pickup is generally reserved for the weeks where we're feeling a bit short on time.
Really cool to see things still being built in Haskell! How do you find using it compared to some of the newer languages that have more modern tooling?
Did you implement your own OAUTH2 flow in haskell for this?
For me, Haskell is the language of 2026. Having an agent available if you get stuck with some weird type error is a blessing. It also helps with the tooling. Though the modern tooling with cabal is pretty good.
I want to use this but I literally have a Rewe 20 meters (at most) from the entrance to my building. Still I might give it a go considering all the positive comments in the thread.
been ordering from REWE for years =) always wished the app was less annoying. all my programm skills - for what? lol. probs for reverse engineering the mTLS part, most people would have just given up.
how stable this is in practice? Internal APIs tend to break without warning when they update the app.
85 comments
I wrote a skill some time ago to support me with "agentic groceries" on my own - it's the future of shopping I would say.
My workflow:
- I paste in urls or text for receipts I will cook this week - agent extracts the ingredients, calculates cups to ml and such, replaces meat with vegi ingredients, replaces some other things I prefer always (often also creates a nice markdown receipe at this steo I can put into Obsidian) - check my list of favs, check search cache (so not every time the api is called, I'm a good netizen :D ) - ask me which items I have at home (no need to add to the basket) - search rewe api for multiple candidates and let me choose. - after each recipe I enter /new to start with fresh context - also I have a list of things I buy every week
I still put everything manually in the basket in the end, but this is not the thing which is time intensive.
Also there already exists this reverse engineered project: https://github.com/ByteSizedMarius/rewerse-engineering/
I do have a suggestion for your app though: Have it compare your basket of goods across different markets in your region to show you the cheapest option. I'm pretty sure this possibility is actually one of the reasons they locked down the API.
I've used Data from REWE in the past and made a comparison between a couple of cities in Germany (I believe it was Frankfurt, cologne, Berlin, Munich and Hamburg). Hamburg was by far the most expensive, often as much as 10-20% more expensive.
I really like your suggestion. I will put it in an issue and look into that. https://github.com/yannick-cw/korb/issues/4
> I do have a suggestion for your app though: Have it compare your basket of goods across different markets in your region to show you the cheapest option.
This is a great idea. I just think the use case is not that big since REWE is the worst in the price/quality ration and just going to another shop would save you more.
>I do have a suggestion for your app though: Have it compare your basket of goods across different markets in your region to show you the cheapest option.
I'd settle for just being able to sort items by unit price... I'm sure this is a [regulation-]solved problem in Germany though
> I'd settle for just being able to sort items by unit price
What do you mean? The official REWE app and website provide just that.
> I'm sure this is a [regulation-]solved problem in Germany though
Not sure what you mean by that.
As I was writing it, I realized that this kind of tactic just feels like it would be banned in Germany
I just checked and REWE only lets you sort by absolute price. But honestly, you can compare prices so much better on their website than in a physical supermarket already [0].
[0] https://www.rewe.de/shop/c/frisches-obst/?sorting=PRICE_DESC You have to enter a random zip code eg 20249
> Compare process across different markets.
Check out smhaggle app on Android
https://play.google.com/store/apps/details?id=com.smhaggle.a...
What I suspect though: They mainly show current discounts. The REWE API exposes those as a separate list for each market. There's around 3.5k markets and each can set their own discounts and has their own product catalogue with their own pricing.
So it would be 3.5k API calls to fetch all offers for all markets. Which is doable.
But fetching all products takes like 100 calls per market. It's quite a bit of data. And I think most supermarket don't publish their catalogue at all since they don't have delivery options.
This is the future of doing groceries. Let us login with our credentials and let us do the search/filling the cart with agents.
Totally fine to do the payment only on the web, so everyone can be sure they only order what was wished, and not 300 avocados.
I think a big issue here is the lack of standard - there is no established way of where an MCP server should be hosted so that agents are easily able to find it. Right now, the best solution I could think of would be to serve it at something like rewe.de/shop/mcp and you'd manually have to register it with your agent.
It can search for items, add them to the basket, picks a delivery slot and does the checkout.
With a little more scaffolding in markdown files, this now takes care of my weekly shopping.
The basic idea is to write a prove in Lean4 and then test both the production implementation (Haskell) and the Lean implementation against random inputs. Compare if the results are the same.
If that is the case -> you can be pretty sure the unproven production version is as correct as the proven lean version.
https://www.dev-log.me/formal_verification_in_any_language_f...
isRegional. That's maybe close enough.B2C: Is it really surprising that a busines has no interest in providing more price transparency to their customers?
The owners of German supermarket and car companies are really the richest of the rich in Germany (okay and maybe the SAP guy on top). It would definitely be a net positive if someone manages to scrape and compare their prices.
In the restaurant market it's one player abusing many small players.
And honestly, I think the reason everyone cries when "Amazon launches an API" is because Amazon would not dare to piss off the German supermarket oligopoly.
> It would definitely be a net positive if someone manages to scrape and compare their prices.
There's a few projects doing that for DE / AT at least.
The issue is that each market sets their own prices and I believe REWE is the only large one where you can fairly easily scrape the product catalogue. I thought about it in a shopping list context, so you'd need to make it location dependent to be useful. But you could do a lot of cool things with it. Like choose a basket of goods and it creates a route for you: "Go to supermarket A and buy goods XYZ, then to supermarket B and buy ABC"
https://www.supermarkt.at https://preisrunter.at https://sparpionier.com
> B2C: Is it really surprising that a busines has no interest in providing more price transparency to their customers?
Might I suggest you remove your tin-foil hat and consider that:
If you think about it, imagine if REWE officially exposed an API B2C. This would mean they are obligated to provide support.Do you really want the price of your shopping to increase because REWE now needs to find money to pay for a helpdesk for the millions of B2C API users ?
Businesses and services differentiating between B2C and B2B is nothing new, that is why the two different terminologies exist !
What next, you don't want to fill up your car at the petrol station (B2C) but you want to be permitted to buy a barrel crude oil direct from the drill and refine it yourself (B2B) ?
> Might I suggest you remove your tin-foil hat and consider that:
First up: Read and follow the rules. No need to insult me. Especially considering what you said shows that you both misunderstood AND misrepresented what I've said.
And frankly, my reasoning was simply saying "Company won't publicize internal info if they don't get an advantage from doing so". It's literally the same reason Google doesn't publish all of their source code. I'm struggling to see what part you are misunderstanding but it has to be something extremely basic to conclude I'm a conspiracy nut for basically stating "Company acts in their interest".
Opening an API to the public allows third parties to develop apps that can then be consumed by end-consumers. Not trying to be offensive here, but do you know what an API is? To conclude I meant every single end-consumers building their own app is at best disingenuously twisting my words.
Opening the API would allow new players like you and me to enter the market and take a piece of the pie. Why would a market, dominated and controlled by a few big players, opt for that? You don't even need to know that the German grocery market is incredibly price competitive, to understand that.
> If you think about it, imagine if REWE officially exposed an API B2C. This would mean they are obligated to provide support. Can you provide a source for that requirement? I'm pretty sure you just made that up.
> Businesses and services differentiating between B2C and B2B is nothing new, that is why the two different terminologies exist ! At this point I'm entirely lost what you read in my comment. Yes I know. I specifically made that distinction.
> What next, you don't want to fill up your car at the petrol station (B2C) but you want to be permitted to buy a barrel crude oil direct from the drill and refine it yourself (B2B) ? Yeah you definitely misunderstood something... What I said/meant:
The question: Why isn't the API open?
My answer: For B2B I gave an example where the API is used by another German firm, providing an example that the API is indeed consumed B2B.
For B2C: They have no reason to do so. They have a well functioning app where you can order stuff. They have one of the bigger recipe pages (at least it does very well SEO-wise) in Germany where you can immediately order ingredients from a recipe. The biggest recipe page in Germany (chefkoch) offers a direct link from recipes to their order page. Maybe you're missing this info? Thinking it's an internal API to data that isn't exposed anywhere at all would somehow explain whatever you tried to say here. But again, if you're that uninformed, don't insult people.
> Opening an API to the public allows third parties to develop apps that can then be consumed by end-consumers. Not trying to be offensive here, but do you know what an API is? To conclude I meant every single end-consumers building their own app is at best disingenuously twisting my words.
Here you are wrong too.
If you want to develop an app via an API that is only offered B2B, what do you do ?
Yes, that's right ...
You phone up REWE and negotiate a license to access to the B2B API to develop your application. B2B2C if you want to put it in simpler terms.
My original point stands. REWE clearly do not want to officially expose the API B2C, almost certainly for the exact reasons I already spelled out in my original post.
But no, its easier for you just to spread FUD, claiming "busines has no interest in providing more price transparency to their customers" just because they will not let you have access to the API as direct B2C.
What you describe as B2B2C is exactly what chefkoch does. And it's exactly what I initially said, so I'm not sure what point you're trying to make. But anyway. Doesn't feel like we're getting anywhere. Have a great day ;)
Haskell is indeed an interesting choice. ;)
> Finally the best side projects are the ones you actually use and this one will be used for all my future grocery shopping.
Until it breaks in a few weeks.
It would have been a cool project!
The workflow in our house is basically to use Siri to add items to an iOS Reminders list as the week goes on. Then, the day of or day before we plan to shop we fill-in with commonly purchased items or one-off things and go shop.
It's been on my TODO list to have an agent "skill" take the Reminders list, ask for additional items, and populate the cart to schedule a pickup. We tend to prefer going in the store though to browse the produce because our local Kroger isn't great about that. Pickup is generally reserved for the weeks where we're feeling a bit short on time.
[1] https://github.com/CupOfOwls/kroger-api
Did you implement your own OAUTH2 flow in haskell for this?
how stable this is in practice? Internal APIs tend to break without warning when they update the app.