"Have an xBloom? Check this before switching beans": a Saudi-made recipe tool for a niche coffee machine

Rashf (رشف, Arabic for "take a sip") is a free tool built by an indie developer in Saudi Arabia. Snap a photo of a coffee bean bag, and it picks the closest of 5,529 public xBloom recipes, blends a brew for that bag, and saves it to the xBloom App or sends it straight to the machine in one tap. The website launched in October; in one month, its creator's TikTok account gathered over 120 million views, and the iOS app came out in October 28. It's small and doesn't make money, but it does a clean job of combining "hardware companion tool + short-video audience finding." Below: what it is, where its users come from, why it worked, and whether you can follow it.

A video that got more shares than likes

2026 年 10 月 24 日,TikTok 账号 @rashf.net 发了一条视频,文案是沙特口语:「عندك xBloom؟ لا تحط محصول جديد قبل تشوف هذا」——「你有 xBloom 吗?换新豆子之前先看这个」,后面一句「拍一下豆袋,用 رشف 拿一份为这包豆子做的配方」。

As of November 6, that video had 236k views, 5,898 likes, 5,210 shares and 6,504 saves (TikTok, 11-06). Shares outnumbering likes is rare on TikTok: viewers didn't just tap like and move on, they sent it to a friend who had just bought an xBloom. The previous few videos ended by asking directly: "Save this and send it to someone who just bought an xBloom."

The next day, October 25, it posted another one: "Why does the café's coffee taste better than yours? The barista dials in a separate recipe for every bag... رشف does it for you." It got 258k views, 4,386 shares and 7,306 saves. It remains the account's top video to date.

What it is

The xBloom is a programmable pour-over coffee machine costing about $500: grind, water temperature and every pour step can be written as a recipe, saved, and the machine brews to it. The problem is finding the recipe. Saudi specialty coffee lovers are numerous. Roasters often include an xBloom recipe with a bag of beans, but they're scattered across different websites, Instagram and Telegram groups. Every time you switch to a new bag, you have to search again. If you can't find one, you experiment yourself, and "the first cup is always sour or bitter." That's the pain point Rashf's videos keep hitting.

Here's what Rashf does: you photograph the bag, and it reads the bean name, roaster, origin, processing method and flavor notes from the label. You choose hot or iced and your taste preference. Then, from the 5,529 public recipes it has collected (from 326 roasters), it picks the closest 5, weights them by similarity, and blends a single recipe: dose, grind size, water temperature, and how much to pour at each step. After tasting, you can tell it "too sour," "too bitter" or "too weak," and it will adjust the recipe. The final step requires no manual copying: one tap saves it to the official xBloom App, and since version 29 in October 1.2.0 it can also send the recipe to the machine over Bluetooth.

Its website lays out the rules in full: bean name weight 28, processing method 12, flavor 12, roaster 10, hot or iced 10, origin 9... Every number is tagged with its source. Iced recipes grind 4 steps finer than hot ones, because 706 iced recipes were compared against 3,198 hot ones. The page ends with: "This integration is unofficial and depends on xBloom's current App interface." For a group of heavy enthusiasts who will spend half a day researching a single bag, that transparency is itself a selling point.

The app is free, with no in-app purchases. App Store reviews: "This app is more useful than xBloom's own app, because I can pick coffee without spending money" and "It turned me into a brewing pro" (Saudi store, 10-29, 10-30).

How big it really is

Tiny.

Category charts in small Gulf countries have low thresholds, so these rankings don't mean many downloads. But look at the xBloom market as a whole: the xBloom official app has only 81 ratings in the US store, averaging 2.86, and 21 ratings in the Saudi store (App Store, 11-06). The installed base of the machine itself is small, so a third-party tool that's been live for 9 days and has 13 ratings isn't a small number.

Where users come from

Almost entirely from its own TikTok account. The first video was in October 8. The high-view videos all came before the app launched: 10-18's "Have an xBloom? Don't do these three things" with 231k, 10-20's with 165k, and 10-24 and 10-25 each with over 200k. The formula is fixed: the opening line "عندك xBloom؟" (Have an xBloom?) filters the target audience, the middle is a brewing tip, and the ending asks you to save it and send it to someone who just bought the machine, with the product name and URL on the last line. Saves and shares in the thousands look like organic spread; from the outside, there's no way to tell whether paid promotion was involved.

Website first, app second. rashf.net works without logging in, and throughout October its traffic went to the website. On 29, it posted "رشف finally has an app, search رشف in the store" (15k views), and app ratings started appearing the next day. There was a catch: "رشف" is a common name in the Arabic app stores (coffee shops, delivery apps and games all use it), so it isn't always easy to find by search.

Riding on two years of xBloom's installed base in the Gulf. xBloom entered Saudi Arabia through the distributor Specialty Hub in September 2024. Searching "xbloom وصفة" (xBloom recipe) on TikTok turns up a large number of coffee creators in Saudi Arabia and the Gulf, with many recipe videos that have hundreds of thousands of views. These people are constantly asking "how do I brew this bag?", so Rashf didn't have to educate the market. No creator was found promoting it.

Other channels are basically absent. No one mentioned it on X for 30 days, and there is no media coverage.

After the app launched, view counts clearly dropped: after 11-03, single videos only reached a few hundred to four thousand. The account also began changing its message, from "Have an xBloom?" to "V60? AeroPress? Any of them," trying to widen the market to all pour-over users.

Why this one

The feature wasn't invented by it. In the English market, Fellow's Aiden Profiler already did exactly the same: photograph the bag, produce a recipe based on your taste, serving a different programmable coffee machine, the Fellow Aiden. Rashf's page structure and copy are very similar. Companion tools for xBloom aren't new either: there's an official recipe library and RFID recipe cards, two Arabic recipe-sharing sites exist in Saudi Arabia, Ratio (163 recipes) and كاف (492 recipes), and in English there are xRecipe and XBRW++. But these are either just a library you have to browse yourself, or English-language and aimed at geeks.

Rashf took off because of several things combined:

1. First in the Arabic xBloom community to do "a recipe from the bag of beans in your hand." Others offered piles of recipes; it offers one answer.
2. The data was already there. Saudi roasters were already in the habit of attaching xBloom recipes to bags of beans. It gathered these scattered recipes, which gave it a base of over 5,000.
3. It finished the last step. Using the xBloom App's interface to save recipes and pushing them over Bluetooth, users go from photo to coffee without touching a parameter.
4. TikTok found the right people. A single line, "Have an xBloom?", filters the audience. In a market this small, 1.2M views a month can nearly reach every xBloom user in the Gulf (inferred).
5. Free, no login, Saudi colloquial Arabic.

Weak points

• The ceiling is how many xBloom machines have been sold in the Gulf. It now wants to expand to V60 and AeroPress, but pour-over recipes are available everywhere online, and those markets don't have the pain point of "scattered data and a hard last step."
• It doesn't make money. It's free with no in-app purchases, and there's no sign of partnerships with roasters or distributors.
• Its survival depends on someone else. Saving recipes and sending them to the machine both rely on xBloom's unofficial interfaces. If the official side changes its interface, or builds "scan bean, find recipe" itself (the official app already has an early version of bean scanning and recipe lookup), its most useful step disappears.
• Growth is slowing. Since the app launched, growth in video views and ratings has been declining.

If it were you

Grade C. You can learn from it, but copying it isn't worth the effort. The idea has already been done by it in the Arabic xBloom community, the market is small, there's no revenue, and its critical point depends on someone else's interface. What's worth learning is the model and the playbook: build a companion tool for a "programmable but scattered-content" device, then use short videos that filter the audience with one line to cold-start.

Who it suits: indie developers who are heavy players in a hardware community, can build a small tool combining photo recognition with rule-based blending, and can make short videos. Treat it as a project to build a portfolio and a following, or a future partnership with roasters and distributors.

Who it doesn't suit: anyone who wants to make a living from it, or anyone outside that community who doesn't know what players are constantly asking about.

How to do it: Don't just build another Arabic xBloom clone. Directions worth testing:

• Switch the device. Find another machine with recipes, parameters, and a community, but scattered content: smart ovens, bread makers, espresso pressure profiles, programmable grinders. First check whether something like the Fellow Aiden Profiler already works in the English market.
• Switch the language. Check whether xBloom or Fellow Aiden has a dense player community and roaster recipes in Japan, Korea, or the Chinese-speaking world. If so, follow what Rashf did.
• Monetize what they haven't. Partner with roasters: scan the beans to generate a recipe, then promote the same roaster's other beans, with revenue split per order.

How to get users: Copy its TikTok playbook. Open by naming the device ("Do you have an XX?"), give one genuinely useful tip in the middle, and end by asking people to save it and send it to a friend who just bought one. Launch a no-login web version first, build a user base, then use a video to funnel people to the app.

When to move: No rush. This kind of opportunity doesn't depend on a time window. It depends on whether anyone has built for the device you picked yet. Check that before you start.

Signs you've got it wrong: ① The official app for your chosen device adds "scan beans to generate a recipe" itself, or blocks third-party access. ② After two months, your videos take off, but ratings and retention stall at a few dozen, which means the niche is too small to support a product. ③ The reverse: if Rashf's ratings in Saudi Arabia passed a few hundred by early 2027, or it signed roaster partnerships, that suggests small-niche tools can grow bigger than they look and are worth reassessing.

Data and sources

• App Store: apps.apple.com/app/id6825658485 (description, reviews, Saudi store); ratings and Food & Drink category charts across stores, 10-30 to 11-06; official xBloom app ratings in the US and Saudi stores, 11-06.
• Website: rashf.net, rashf.net/en/about (recipe count, roaster count, matching weights, API notes).
• TikTok: @rashf.net videos with plays, likes, shares, and saves, 11-06; search "xbloom وصفة".
• xBloom enters Saudi Arabia: article on the xbloom.com website, 2024-09-10.
• Competitors: fellowaidenprofiler.com, xbloomratio.com, caf-xbloom.com, xbloom-app.vercel.app, collective.xbloom.com, and XBRW++ on the App Store.
• Places marked "inferred" or "unverified" are inferences or self-reported. No public figures exist for downloads, revenue, or paid acquisition.