Translate an Amazon Listing to German With AI: Keywords and Checks

To translate an Amazon listing into German with AI, give the model your English title, bullets and description plus the product facts, ask for German written for shoppers rather than word for word, then build your German keywords from Amazon.de's own search suggestions. Before publishing, have a native reader check the wording and every claim you can be held to.

Translation is the easy 20 per cent. The listing sells or does not sell on the words German shoppers actually type, on whether your claims are true and supportable in the form you wrote them, and on small details such as units and sizes. A fluent draft that uses the wrong shopper word, or promises "waterproof" where the test only supports "water-repellent", can cost you sales or get the listing suppressed.

Follow me on Instagram@sagnikteaches

What you are translating, and what you are rewriting

Treat the listing as five separate jobs, because each has a different rule.

Connect on LinkedInSagnik Bhattacharya

Amazon's Build International Listings tool cross-lists products into other Amazon stores, and some sellers let it, or a paid translation service, produce the first German text. Whatever produced the draft, check it exactly as you would an AI draft, line by line, before it goes live. For the English side of the listing, this tutorial on bullets and search terms covers the groundwork.

Subscribe on YouTube@sagnikteaches

The limits that shape the German title and search terms

Two limits matter before you write a word. From 27 July 2026 Amazon caps product titles at 75 characters, spaces included, in all categories except media, and a German retailers' association reports the same limit for Amazon.de. A separate Item Highlights field adds 125 characters for materials or uses. Titles still over the limit are gradually replaced with Amazon's AI recommendation (brand owners get 14 days to review it), which you do not want for a brand you care about. German words are long, and a literal translation of an English title almost always exceeds 75 characters.

The backend search terms field is counted in bytes, not characters: the limit is 250 bytes, with one byte per basic letter or digit, and Amazon staff note that characters such as ä can take two or more bytes each. A German list full of umlauts and ß therefore holds a little less than the same list in English. Check the counter in Seller Central for your store when you paste the terms in; the limits and counting rules change, and the screen is the authority.

A main example: translating a waterproof jacket listing

This is an illustration with invented specifications. An online clothing store sells a women's lightweight waterproof jacket in green. The English listing already works well. They want an Amazon.de version. Here is the English title and bullets.

Title: [Brand] Women's Waterproof Rain Jacket, Lightweight Packable Hooded Coat, Breathable, Green
Bullet 1: FULLY WATERPROOF - 10,000 mm water column and taped seams keep you dry in heavy rain
Bullet 2: BREATHABLE - mesh lining lets moisture escape on walks and bike rides
Bullet 3: PACKS INTO ITS OWN POCKET - weighs 320 g, fits in a handbag
Bullet 4: ADJUSTABLE HOOD AND CUFFS - keeps wind and rain out
Bullet 5: MACHINE WASHABLE - 30 degrees, no tumble dry

The facts a translator needs are the spec sheet, not the English copy alone: the test result behind "10,000 mm", the fabric composition, the weight, the sizes in centimetres and the care label. Give the model all of it.

Step 1: ask for a translation that keeps the facts

Translate the Amazon listing below into German for Amazon.de shoppers. Use "Sie" in any direct address. Write for meaning, not word for word. Do not add any claim that is not in the facts. Keep numbers and units exact (convert nothing yourself). If an English term has no standard German shopper word, tell me in a separate list called "Check with a native reader".

Facts: 100% polyester shell, PU coating, water column 10,000 mm (tested by an outside lab, report available), taped seams, weight 320 g in size M, care 30 degrees, no tumble dry.

[paste listing]

A realistic result (illustrative) for bullet 1 is: "VOLLSTÄNDIG WASSERDICHT: 10.000 mm Wassersäule und verklebte Nähte halten Sie bei starkem Regen trocken". The German uses a full stop for the thousands separator, which is right for German text. The note at the end of the output lists three items to check, such as whether "verklebte Nähte" is the shopper phrase for taped seams, and how to render "packable". Those flags are the most useful part of the answer, because they point the human reviewer at the doubtful words.

The title is where literal translation hurts. The model might return something like "[Marke] Damen wasserdichte Regenjacke, leichte, packbare Kapuzenjacke mit Kapuze, atmungsaktiver Mantel, Grün", which is 109 characters and repeats "Kapuze". It needs rewriting for the limit and the search term. A better German title (illustrative, to be confirmed by research) is "[Marke] Regenjacke Damen wasserdicht, leicht, packbar, mit Kapuze, Grün", which counts 71 characters with the placeholder; recount once your real brand name replaces [Marke]. It leads with the product word and gender, which German shoppers commonly type, and drops the repeated hood.

German keyword research, not keyword translation

Take the English phrase "rain jacket women". A translator gives "Regenjacke Damen", which happens to be a natural phrase. Now take "packable". There may be no single common word, and shoppers may search for "faltbar", "packbar" or "leichte Regenjacke zum Verstauen". If you translate your English keywords one by one, you will miss those. The point of German research is to find what people actually type, which AI cannot know.

A practical method, with no paid tool needed:

1. Open Amazon.de and type your seed words into the search bar: "Regenjacke", "Regenjacke Damen", "Regenmantel". Write down every autocomplete suggestion. These come from real searches.
2. Open the top ten results for your main phrase and copy their titles and bullets into a sheet. Note recurring words, pack sizes, materials and features they name.
3. Ask AI to cluster the list, not to invent it (prompt below).
4. Decide the title phrase, the bullet words and the backend terms from the clusters. Keep competitors' brand names out.
5. Check that each backend term adds something the title and bullets do not already say.

Below are German phrases collected from Amazon.de autocomplete and from the titles of competing listings for women's rain jackets. Group them into clusters by meaning (product type, feature, material, use, colour). For each cluster, name the most common wording. List terms that look like a different product (for example a long coat or a poncho). Do not add any phrase that is not in my list.

[pasted phrases]

An illustrative result separates "Regenjacke" and "Regenmantel" as two different products. A jacket ends at the hip, while a Mantel is long. That tells the store which term suits its jacket in the title, and which term belongs only in a backend term because shoppers do sometimes use it for shorter jackets. A model cannot give you search volume, and it should not pretend to. If you want volume numbers, use a keyword tool or Amazon's own brand analytics where you have access, and check that tool's pricing and coverage yourself.

The backend field might then hold something like "regenmantel wanderjacke outdoorjacke fahrradjacke überjacke regencape windjacke trekkingjacke wetterjacke", which is 106 bytes against the 250 limit (counting the ü as two). That leaves room, though more terms are not always better: only add words that describe the product accurately.

Checking claims before they go live in German

Claims are where a translation can move from clumsy to costly. English and German do not carve up claim words the same way, and consumer and marketplace rules treat some claims strictly. The principle is simple: say only what you can document, and use the word your document supports.

• Waterproof versus water-repellent. German has distinct words, "wasserdicht" and "wasserabweisend", and they are not interchangeable. If your test supports a given water column, say that figure, and use "wasserdicht" only where your evidence matches how the term is understood for clothing. If you are unsure, say "wasserabweisend" or give the measured value alone.
• Organic, eco and green. Words like "Bio", "nachhaltig" or "umweltfreundlich" are heavily scrutinised. Use them only with a certificate or specific, checkable facts ("enthält 60 % recyceltes Polyester" if it does and you have the supplier's proof).
• Health and safety. Anything that sounds medical, antibacterial, anti-allergy or protective needs support. In food and supplements, health claims for sale to customers in the EU are tightly controlled, so check the wording you plan to use with an adviser or the marketplace's category rules.
• Certifications. Name a certificate only if it covers this product and is current. Keep the certificate number on file.
• Comparisons and superlatives. "Best", "number one" and "the only" translate easily and are hard to defend. Remove them unless you can prove them.

Also check whether the product category has compliance requirements for the German store, such as product safety documents, labelling or packaging registration. Amazon's compliance pages in Seller Central and your adviser are the authority, so confirm what applies to your category, including any packaging registration, before you list. The tutorial on AI listing mistakes covers how a smooth draft can mislead without anyone meaning it to.

The native-reader check

A native reader does three jobs that AI cannot do for you: tell you what sounds wrong, tell you what sounds odd, and tell you what a shopper would type instead. Hire a freelancer, ask a German-speaking friend or ask a translator for a "review and edit" of a machine draft, usually quicker and cheaper than translating from scratch. Ask for a quote with a fixed word count, and give them the facts sheet as well as the listing.

Send a short brief rather than the file alone.

This is a machine-translated draft of an Amazon.de listing for a women's rain jacket. Please:
1. Mark anything that sounds unnatural to a German shopper, and suggest a replacement.
2. Tell us if any claim (waterproof, breathable, 10.000 mm) would read as more than the facts support. The facts sheet is attached.
3. Say which search words you would expect a shopper to type for this product, from your own use of Amazon.de.
4. Check units, sizes and the tone (Sie, not du).
Please mark changes so we can see them.

When the reviewed text comes back, compare it with your AI draft. The differences teach you what to put in your next prompt: if the reviewer always changes "packbar" to another phrase, add that to the prompt as a glossary entry. Over three or four listings your drafts need fewer edits, and the reviewer's time falls.

Before and after: three bullets, edited

To see what review catches, compare an illustrative draft with the edited version.

Bullet 2, draft: "ATMUNGSAKTIV: Netzfutter lässt Feuchtigkeit entweichen bei Spaziergängen und Radtouren". Reviewer's note: "Netzfutter" reads like a fishing net; "Mesh-Innenfutter" is commonly understood. Edited: "ATMUNGSAKTIV: Das Mesh-Innenfutter lässt Feuchtigkeit entweichen, beim Spaziergang wie auf dem Rad".

Bullet 3, draft: "PACKT SICH IN IHRE EIGENE TASCHE: 320 g Gewicht, passt in eine Handtasche". Reviewer's note: ambiguous; it sounds as if the jacket packs into your pocket. Edited: "PLATZSPAREND: Wiegt nur 320 g (Größe M) und lässt sich in der integrierten Tasche verstauen".

Bullet 5, draft: "MASCHINENWASCHBAR: 30 Grad, kein Trockner". Reviewer's note: add the degree sign and fix the care instruction to match the label. Edited: "PFLEGELEICHT: Maschinenwäsche bei 30 °C, nicht im Trockner trocknen".

None of these edits is a big idea. Together they are the difference between a listing that reads like a translation and one that reads like a shop.

Units, sizes and the details that follow translation

Several details sit outside the text but ruin a listing if missed.

• Units. Use metric: centimetres, kilograms, litres, degrees Celsius. Convert from the exact measurement, not from a rounded English figure, and write the number the German way, with a comma as the decimal mark (for example 0,5 kg).
• Sizes. Letter sizes and numeric sizes do not map one to one across markets. Build the size chart from body measurements and compare it with German competitors' charts. If your garment runs small, say so in the bullets.
• Colour names. "Forest green" may become "Waldgrün", but the filter colour in the listing fields must use Amazon's own options.
• Images. Redo infographics, lifestyle overlays and the size chart in German. English text on any image weakens trust, so check Amazon's image requirements for your category before you add text to the main image at all.
• Legal and brand. Product safety labels, manuals and inserts may need German text, and some categories need specific information in the listing. Check Seller Central's help for your category.
• Price and delivery. Set the price in euros after working out fees and costs; a conversion at today's rate is not a pricing strategy. The FBA versus FBM cost model tutorial shows a way to check margins by store.

Another product, another trap: a furniture listing

A furniture maker listing a solid oak side table meets a different set of problems. German has precise words for finishes and woods, and compound nouns carry meaning: "Beistelltisch Eiche massiv" says more than "oak side table". AI may produce "Eichenholz Seitentisch", which is understood but not what shoppers type. Dimensions need care: width, depth and height in centimetres in the order shoppers expect, and a clear statement of whether the table arrives assembled. Claims such as "FSC-zertifiziert" need a current certificate, and "Massivholz" means solid wood throughout, so do not use it if the top is veneered. Here the native reader's checks run through materials and assembly, not clothing performance, but the pattern repeats: draft with AI, research the shopper's word, verify claims, review with a native speaker.

Build a small glossary so the next listing starts ahead

After the first review, you hold a valuable thing: a list of German words that a native reader approved for your products. Keep it in a sheet with four columns: English term, approved German term, words to avoid, and a note on why. Paste the sheet into every later translation prompt and tell the model to use it without exception.

Use this glossary for every translation. If a term appears in the glossary, use the approved German term exactly. Never use a term from the "avoid" column. If you need a term that is not in the glossary, add it to a list called "New terms" at the end.

English | Approved German | Avoid | Note
rain jacket | Regenjacke | Regenmantel | Mantel means a long coat
packable | platzsparend verstaubar | verpackbar | Reviewer's choice
mesh lining | Mesh-Innenfutter | Netzfutter | Netzfutter reads as a net
water column | Wassersäule | Wasserdruck | State the figure with mm

An illustrative result for a second jacket in navy is a draft with all four terms already right and a "New terms" list of two items, such as "adjustable cuffs", which the reviewer then settles in a minute. Over ten listings the "New terms" list shrinks, and so does the review bill. A furniture or pest-control catalogue works the same way: the glossary is the asset, and the model just applies it.

A small testing routine after launch

The listing is not finished when it goes live. Over the first weeks, look at what the store gives you: impressions, clicks, conversion and search-term data if you have access to it. If impressions are low, the title and backend terms may not match what people type; if clicks are fine but conversion is poor, check the images, price and the bullets. Change one thing at a time and write down the date, or you will not know what worked.

Read the first German customer questions and reviews with care. Questions often show what the listing failed to say, and a customer who says the sizes ran small is a fact to put into the next edit. Ask AI to summarise ten German reviews into themes if you cannot read them yourself, then check the summary against a reviewer. The small seller tutorial shows other listing and review jobs that fit this stage.

Cost, time and when to skip the shortcut

A single listing takes about an hour of your time with AI: ten minutes collecting facts, twenty on research, ten on the draft prompts, and twenty comparing and adjusting. The native reader adds a day or two of waiting and a modest fee, which you should get quoted before you commit. Spread across a catalogue, build a shared glossary of your product terms in German once, then reuse it in every prompt.

Skip the shortcut when the product is regulated heavily, such as supplements, cosmetics, children's goods or electrical items. There, the listing text is the smaller part of the job and the documents matter more; get advice first. Where the cost of a bad listing is higher than a day's review, the review should not be optional. For broader translation choices, the tutorial on machine translation for client-facing documents sets out when machine output is good enough on its own, and when a person must sign it off.

Checking it worked

Work down a short list before you publish. Does the title fit the limit and lead with the German product word? Do the backend terms fit the byte counter and avoid repeating the title? Is every claim backed by a document in a folder? Did a native reader see the final text, not only the first draft? Do the images, units and size chart match the bullets? If all answers are yes, publish and keep the folder: the checklist, the reviewer's notes and the facts sheet are the start of your German glossary for the next listing.

Questions sellers ask about German Amazon listings

Can I just let Amazon translate my listing?

Amazon's Build International Listings tool copies offers into other Amazon stores, and any German text it or a translation service fills in for you is a machine draft. Treat it as unreviewed: check the keywords, units, claims and tone yourself or with a native reader before it carries your brand name.

Do I need a native German speaker?

Yes for the check, though not for the whole translation. AI can produce a serviceable draft, but a native reader catches awkward phrasing, wrong shopper words and claims that sound legal in English and risky in German. An hour or two of review is a sensible cost for a product you plan to sell for years.

Can I reuse my English search terms?

Not as they are. German shoppers use compound words and different product names, so translated English keywords often match nothing they actually type. Build the German list from Amazon.de's own search suggestions and competitor listings, then fit it within the byte limit for search terms.

Which claims are risky?

Health, safety, environmental and certification claims, such as waterproof ratings, organic or eco labels, and any word like antibacterial or medical. Only claim what you can document, and use the wording your test report or certificate supports. When unsure, leave the claim out.

Should I translate the images too?

Yes. English text on images makes a listing look foreign and can hurt trust. Redo any infographic or size chart in German with metric units, and check the translated text against the bullets so both say the same thing.

Further reads

• How to Use AI for Etsy Titles, Tags and Descriptions — Titles and tags thinking carried over to another marketplace.
• Amazon Seller Virtual Assistant or AI: Which Tasks Go to Whom — Which listing tasks go to a person and which to AI.
• Translate Your Shopify Store: Translate & Adapt vs DeepL vs ChatGPT — Translate your own store as well as Amazon.
• Walmart Marketplace vs Amazon: Which Should a Small Seller Add Next? — Compare adding another marketplace with adding a language.
• AI for Amazon Listing Images: Infographics Inside Amazon's Rules — Plan, draft and check Amazon listing infographics with AI while the main image, text overlays and claims stay inside Amazon's rules.
• Amazon's 75-Character Title Rule: Review and Override AI Rewrites — Find the titles Amazon's AI has rewritten, check them against a short rubric, and put your own 75-character version live.
• AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.

Sources: Amazon's Seller Forums announcement of the 75-character title limit and Item Highlights (updated 10 June 2026), a retailers' association notice on the same limit for Amazon.de, Amazon staff replies on byte counting in the Seller Central forums and Linnworks' Amazon search-terms help (250 bytes), all checked November 2026; always check the current rules in your Seller Central account.