How to scrape food and grocery data from Linguee using the Minexa.ai extension

Bilingual food and grocery data sitting inside a translation database is not the most obvious scraping target. But if you work in food labelling, cross-border product research, or multilingual content, Linguee is one of the few places where real-world sentence pairs from verified sources are publicly available at scale.

This guide walks through how to extract that data from Linguee (linguee.com.br) using the Minexa.ai Chrome extension, with no code required. Before getting into the steps, here is what the extracted data actually looks like.

What the data looks like

Each row returned by the scraper represents one bilingual sentence pair from the Linguee results page. Below is a sample of two cleaned records from a search for organic food terms:

[
 {
 "row_id": "row_0_2711327943_0",
 "source_website": "natursoy.es",
 "related_websites": "natursoy.es",
 "link_href": "http://www.natursoy.es/continguts/al/about-us-al.php",
 "website_url_2": "http://www.natursoy.es/continguts/pt/sobre-nos-pt.php",
 "left_sentence_class": "sentence left",
 "right_sentence_class": "sentence right2",
 "key_information": "Natursoy S.L. ist ein in Castellterçol (Barcelona) ansässiges Unternehmen...",
 "short_description": "natursoy.es",
 "external_links": ""
 },
 {
 "row_id": "row_0_5115250121_0",
 "source_website": "",
 "related_websites": "eur-lex.europa.eu",
 "link_href": "http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=COM:2002:0306:FIN:DE:PDF",
 "website_url_2": "",
 "left_sentence_class": "sentence left",
 "right_sentence_class": "sentence right2",
 "key_information": "die Ermittlung und den Aufbau von Kompetenz in neuen Bereichen...",
 "short_description": "eur-lex.europa.eu",
 "external_links": "http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=COM:2002:0306:FIN:PT:PDF"
 }
]

A few fields worth understanding before you run this yourself:

The long_description field returns a structured array of typed objects. Each object encodes a tag name, an optional attribute, a type, and a text value. This means the full Portuguese source sentence and its German translation are both captured within a single traversable structure per row, with inline bold markers identifying exactly where the search keyword appears in context. You can reconstruct the full sentence from the array or filter only the bolded match segments depending on what your pipeline needs.

The right_sentence_class field looks like a simple CSS string, but it carries meaningful signal. A value of sentence right2 indicates a standard confirmed translation pair. A value of sentence right2 warn means Linguee has flagged that pair as potentially unreliable. If translation quality matters for your use case, filtering on this field lets you separate high-confidence pairs from uncertain ones without any manual review.

The external_links field and the link_href field are not duplicates. When a source document exists in both German and Portuguese versions, link_href stores the German-language document URL and external_links stores the Portuguese equivalent. When only one version exists, one of the two fields will be empty. This parallel structure lets you track both language editions of a source document within a single row.

Watch the full tutorial first

The video below covers the complete extraction workflow from start to finish. Watch it before going through the steps if you prefer to see the process in motion.

https://www.youtube.com/watch?v=ykE6StzLrMM

How to extract the data step by step

Install the Minexa.ai Chrome extension from the Chrome Web Store, then open your browser and navigate to the Linguee page you want to scrape. For food and grocery terms, the starting URL used here is the Portuguese-German search results page for organic food.

Once the page has fully loaded, open the Minexa.ai extension from your browser toolbar. The extension popup will appear and prompt you to confirm you are on the correct page. Click the confirmation button to proceed.

Minexa.ai will then detect the pagination structure of the page automatically. You will see a summary of what it found, including whether a next page button, infinite scroll, or load more button is present. Review the detection and click Continue.

Next, choose whether to scrape the list page only or to also follow each result link and extract detail page data. For Linguee, the sentence pair data is fully available on the list page, so selecting the single list option is sufficient for most food and grocery research workflows.

Select the simple scraping scenario on the next screen. The advanced option is available if you need custom click workflows, but the standard setup handles Linguee results without any additional configuration.

Minexa.ai will now highlight the full container holding the list of translation pairs on the page. This is the repeating structure it has identified as the data source. Confirm the selection and click to create the scraper.

After the scraper is created, all detected data points are displayed with navigation controls so you can review each column before running the job. Fields like long_description, right_sentence_class, external_links, and website_url_2 will all be visible here.

Once you are satisfied with the detected fields, complete the configuration. The scraper is now saved and ready to run. You will also see the option to set up a recurring schedule if you want to capture Linguee data at regular intervals.

Your scraping job will appear at the top of the jobs list. Click Run to start extraction. The job processes each page and returns results as rows in a structured table.

When the job finishes, the full dataset is available for export. Download it as Excel or JSON depending on how you plan to use it.

What you can do with this data

The combination of source website URLs, bilingual sentence pairs, and translation quality flags makes this dataset useful for several food and grocery research tasks. You can filter to only rows where right_sentence_class does not contain the warn flag to build a clean corpus of verified Portuguese-German food term translations. The related_websites field identifies which domains contributed each sentence pair, so you can assess whether your results skew toward regulatory sources like EUR-Lex or toward commercial food producers.

If you want to get started, install the Minexa.ai extension and run your first extraction in a few minutes: Get the Minexa.ai Chrome extension.

For a related guide on extracting structured data from another content-rich source, see: Scraping Azul Fidelidade flight deals with the Minexa.ai extension.