{"slug": "what-i-learned-building-a-12-language-ai-product-search-for-aliexpress", "title": "What I learned building a 12-language AI product search for AliExpress", "summary": "A developer built OneFindMe, a free multilingual search layer for AliExpress that resolves shopping intent to a canonical product query before translating, rather than translating raw phrases. The system trims descriptive queries to a product noun plus one or two qualifiers, ranks candidates by real order volume with a light penalty for unreviewed listings, and caches query resolution to cut latency. It supports 12 languages and applies locale-specific filtering for shipping thresholds and regional defaults.", "body_md": "AliExpress has one of the largest product catalogs on the planet, but its search\n\nassumes you already know the exact English keyword. If a shopper can only *describe* what they\n\nwant — \"that little stone thing for face massage\" — the results collapse into noise. I spent a\n\nwhile building [OneFindMe](https://onefindme.com/en/), a free multilingual search layer that sits in front of AliExpress,\n\nand these were the lessons that actually moved the needle.\n\nNaive translation is a trap. \"לק ג'ל\" (Hebrew) machine-translates to \"gel polish\", which on\n\nAliExpress surfaces floor lacquer as often as nail products. The fix was to resolve the shopping\n\n*intent* to a canonical product query first, then translate that — not translate the raw phrase.\n\nA small curated keyword map beat the general model for the high-traffic terms, because the model\n\nkept inventing plausible-but-wrong category ids.\n\nLong, descriptive queries rank worse than short ones on the marketplace API. The product noun has\n\nto lead. \"comfortable running shoes for wide feet women\" performs far worse than \"wide running\n\nshoes\". So the pipeline trims to the product noun + one or two qualifiers before it ever hits the\n\nAPI.\n\nThe cheapest listing is almost never the best answer. Sorting candidates by real order volume,\n\nthen lightly penalising listings with no reviews, produced results people actually clicked. Price\n\nis a filter, not a ranking signal.\n\nProduct availability changes hourly, but the translation of \"wireless earbuds\" into a good query\n\ndoes not. Caching the *query resolution* (and letting unknown terms cache their translation on\n\nfirst use, so the second shopper gets an instant answer) cut latency dramatically without serving\n\nstale stock.\n\nThe same query needs different handling per country: shipping thresholds, which categories pay,\n\neven modesty defaults in some regions where the marketplace's own \"relevance\" surfaces things a\n\nshopper did not ask for. Filtering the junk while never filtering a legitimate intent turned out\n\nto be the hardest, most locale-specific part.\n\nIf you want to see the result, [OneFindMe](https://onefindme.com/en/) is live and free — type what you want in any of 12\n\nlanguages and it does the translating, trimming, and ranking described above. Happy to answer\n\nquestions about any of these in the comments.", "url": "https://wpnews.pro/news/what-i-learned-building-a-12-language-ai-product-search-for-aliexpress", "canonical_source": "https://dev.to/ohadfarkash/what-i-learned-building-a-12-language-ai-product-search-for-aliexpress-2n61", "published_at": "2026-09-20 11:16:33+00:00", "updated_at": "2026-09-20 11:54:24.990247+00:00", "lang": "en", "topics": ["ai-search", "natural-language-processing", "ai-products", "ai-tools"], "entities": ["OneFindMe", "AliExpress"], "alternates": {"html": "https://wpnews.pro/news/what-i-learned-building-a-12-language-ai-product-search-for-aliexpress", "markdown": "https://wpnews.pro/news/what-i-learned-building-a-12-language-ai-product-search-for-aliexpress.md", "text": "https://wpnews.pro/news/what-i-learned-building-a-12-language-ai-product-search-for-aliexpress.txt", "jsonld": "https://wpnews.pro/news/what-i-learned-building-a-12-language-ai-product-search-for-aliexpress.jsonld"}}