Tip: paste a table range including its header row. Media and text columns are identified from normalized header semantics and cell content. Common text fields include product_info, product_details, product_attributes, product_specs, product_features, product_benefits, product_name, product_category, brand, material, color, size, selling_points, ingredients, instructions, review and feedback, in addition to ref_text, creative_title/ceative_title, title, text, context and description. Clear IDs, codes, URLs and media-asset fields are not classified as text merely because their names contain a related word; recognized product landing-page columns such as target_url and landing_page remain separate. Headerless TSV text/JSON columns are heuristically detected and URL-only lists are supported. Media URLs embedded in text fields are extracted and tagged with the original header name and media type. There is no fixed per-group media limit; practical capacity depends on browser performance. Translation offers 24 common target languages. Translation first tries Chrome’s built-in on-device Translator API when available, then Google endpoints and a MyMemory fallback. Remote requests are serialized with a 1.5-second minimum interval, duplicate in-flight requests are shared, and providers returning HTTP 429 are cooled down temporarily; availability still depends on browser support, language pair, and downloaded language packs.