Extract Listing Details from Leboncoin and Marktplaats Screenshots
Manually copying a listing's title, price, and location from a French or Dutch classifieds screenshot takes about 2 minutes per ad — this picks all five fields in under 10 seconds, across both markets.
5-10s per screenshot · European € format normalized · Up to 99% accuracy
Five Listing Fields You Can Pick from Any Leboncoin or Marktplaats Screenshot
Leboncoin renders listings as grid cards with bold "Prix" tags; Marktplaats uses denser text blocks with Dutch labels. With Custom Column Extraction, the AI reads each field by its semantic role — the boldest text is the item name, the amount next to "€" is the price — regardless of which platform the screenshot comes from.
The largest bold text on the listing — a card title on Leboncoin, a headline block on Marktplaats.
French format "150 €" or Dutch "€ 1.250,-" — normalized to clean numbers. "Bieden" qualifiers captured alongside.
City and department on Leboncoin; city and postcode radius on Marktplaats.
The listing body text — plain paragraph on both platforms, captured as-is.
Particulier vs Professionnel on Leboncoin; Privé vs Zakelijk on Marktplaats.
Type these as column names in English, French, or Dutch. The AI reads each field by semantic role across both platforms.
Two Countries, Two Layouts, One Price Format — Why Screenshots Beat Scrapers for FR and NL Classifieds
Leboncoin uses a grid-heavy card layout with "Prix" tags, "Vendu" overlays, and a visible category sidebar. Marktplaats is more text-dense with "Bieden" offer-type pricing and Dutch labels. URL-based scrapers either trigger DataDome anti-bot detection on Leboncoin or rely on Marktplaats's undocumented API — both break when the site updates. A screenshot captures what the human sees, regardless of what the backend allows. Here is why that matters.
Why Leboncoin and Marktplaats Break Traditional Extraction
Leboncoin's grid layout buries the listing in visual noise. Category navigation, "Prix" tags, "Vendu" overlays, and sponsored boxes all produce visible text. Traditional OCR cannot distinguish the listing card's title from the category label next to it.
Marktplaats uses "Bieden" instead of fixed prices. Many listings show no numeric price — just "Bieden" (make an offer) or "€ 1.250,-" with Dutch formatting. Pure numeric extraction either fails or outputs the wrong value.
French and Dutch € formats differ completely. Leboncoin writes "1 500 €" (space as thousand separator). Marktplaats writes "€ 1.250,-" (comma decimal, dash for zero cents). One set of column definitions needs to parse both conventions.
How Semantic Extraction Reads Both Platforms with One Column Set
Name the fields — the AI finds values by meaning, not position. Type Item Name, Price, Location, Description, Seller Type. The visual model identifies the boldest card text as the item name, the number next to "€" as the price, and ignores adjacent category navigation and sponsored labels.
"Bieden" pricing is captured with context. When Marktplaats shows "€ 250 Bieden", the AI puts 250.00 in the Price column and "Bieden" in a qualifier field — preserving both the amount and the knowledge that it is negotiable.
Both € formats normalize automatically. "1 500 €" from Leboncoin and "€ 1.250,-" from Marktplaats both land as 1500.00 in the Price column — with the raw text preserved in a parallel column for verification.
From Mixed FR and NL Screenshots to One Consolidated Listing Sheet
Whether you are tracking items listed on both platforms, comparing prices across markets, or archiving your transactions, here is how the extraction workflow handles screenshots from two countries in a single pass.
Upload Screenshots from Both Platforms
Drop in captures from your camera roll or desktop — Leboncoin listing cards with "Prix" badges and "Vendu" overlays alongside Marktplaats detail pages with "Bieden" pricing and "Privé" labels. PNG, JPG, or WebP from any device. Mix them in the same batch — no sorting by platform needed.
Define One Set of Column Names
Type the fields you want — in English, French, or Dutch: Item Name, Price, Location, Description, Seller Type (or Titre, Prix, Localisation, Description, Vendeur; or Titel, Prijs, Locatie, Beschrijving, Verkoperstype). The AI reads each screenshot and fills matching fields — a Leboncoin card yields the Price from its "Prix" tag, a Marktplaats page reads the same field from its label.
Export a Combined FR/NL Spreadsheet
Processing takes 5-10 seconds per screenshot — about 18x faster than opening each image and typing the fields manually. The output is one XLSX or CSV file with all five columns. Prices are already normalized from both French and Dutch formats. Filter by Seller Type to separate professional listings from private sellers across both platforms.
When Leboncoin and Marktplaats Screenshot Extraction Works Best — and When to Be Cautious
When It Works Best
Full-resolution native screenshots of listing detail pages. High-contrast, machine-rendered text from either platform reaches up to 99% accuracy for fields like Item Name and Price.
Mixed Leboncoin and Marktplaats screenshots in one batch. One column definition works across both platforms — the AI adapts to each layout independently.
Clean captures where the listing card or text block is the main focus. When the listing occupies most of the frame, the AI isolates its fields from adjacent navigation and sponsored content.
When to Be Cautious
Screenshots forwarded through messaging apps. WhatsApp and Messenger compress images, reducing accuracy for smaller text like seller type badges. Use native screenshots for best results.
Multiple listing cards in one screenshot. The AI reads the most visually prominent listing. For clean per-item extraction, capture each listing individually.
Leboncoin screenshots with heavy "Vendu" overlay. The "Vendu" badge is semi-transparent — the underlying Item Name and Price are still readable, but accuracy may decrease for fields near the overlay edge.
Frequently Asked Questions
Can I extract the Item Name and Price from a Leboncoin screenshot that already shows a "Vendu" overlay?
Yes. The AI reads the listing data underneath the "Vendu" badge — the overlay is semi-transparent and does not fully obscure the Item Name or Price. The Description column captures the listing body text from the area below the overlay unaffected. For Marktplaats, "Verkocht" badges follow the same semi-transparent pattern and are handled identically.
How does the tool handle Marktplaats listings that show "Bieden" instead of a fixed Price?
When a listing shows "Bieden" (make an offer), the AI extracts the suggested starting price if one is visible and places it in the Price column, while recording "Bieden" in a separate qualifier field. If no numeric price is shown at all — just "Bieden" or "Te koop" — the Price column remains blank and the qualifier column captures the offer type. This gives you clean numeric data for listings that have fixed prices, without losing the distinction between fixed-price and offer-type listings.
Can I process Leboncoin and Marktplaats screenshots together in one batch?
Yes. Define one set of column names — Item Name, Price, Location, Description, Seller Type — and upload screenshots from both platforms into the same batch. The AI reads each screenshot independently: a Leboncoin card with "Prix" in a bold tag and a Marktplaats detail page with "Prijs" in a labeled row are both parsed using the same semantic column definitions. Every screenshot becomes a row in one spreadsheet, and an optional Platform column can be added to tag each row with "Leboncoin" or "Marktplaats" for filtering.
What columns work best for tracking listings across both platforms?
A practical setup uses the five standard fields: Item Name, Price, Location, Description, Seller Type. Add Platform to distinguish Leboncoin from Marktplaats rows. Add Price Qualifier to capture "Bieden", "Prix fixe", "Vaste prijs", or "à débattre" alongside the numeric price. For pricing intelligence, a Computed Column like Price per Region (Price × Location) can analyze geographic pricing differences across France and the Netherlands.
Can I use French or Dutch column names instead of English?
Yes. Column names can be in any language. Type Titre, Prix, Localisation, Description, Vendeur for Leboncoin screenshots, or Titel, Prijs, Locatie, Beschrijving, Verkoperstype for Marktplaats. The AI recognizes the field names by their semantic meaning and maps them to the correct listing data regardless of the language you use. The output spreadsheet uses your chosen column names as headers.
Return to the secondhand marketplace overview — covers Facebook Marketplace, Vinted, Carousell, Leboncoin, Marktplaats, and more.
Deep dives: Batch processing marketplace screenshots into a single spreadsheet · General guide to extracting data from screenshots