Recognize Item Name, Price, and Location from Daangn Market Screenshots
Manually reading each Daangn Market listing and typing the details into a spreadsheet takes about 3 minutes per ad — this recognizes Item Name, Price, and Location from any screenshot in under 10 seconds.
5-10s per screenshot · Up to 99% accuracy on printed text
Three Fields You Can Recognize from Any Daangn Market Screenshot
Daangn Market listing screenshots mix card data with Korean-language app UI — navigation bars, category filters, and chat previews all produce visible text. With Custom Column Extraction, the AI identifies each field by its role in the listing card, not its screen position.
The listing title in Korean — typically the boldest text in the card, e.g. "아이폰 15 프로 팝니다" or "맥북에어 M1".
The asking amount with Korean formatting — ₩12,000 or "12,000원" depending on the app version.
Neighborhood or district — hyperlocal, e.g. "서울시 마포구" or "성남시 분당구" at the card bottom.
Type these as column names. The AI reads each field by card structure across Korean locale variants and app UI versions.
Korean Text, ₩ Currency, and Hyperlocal Locations — Why Daangn Screenshots Need More Than Simple OCR
A Daangn Market screenshot from Seoul shows "아이폰 15 프로" with "₩1,200,000" and "서울시 마포구" — the same card structure needs to read mixed Korean Hangul and English text, Korean currency formatting with comma separators, and town-level location names that follow no single pattern. Here is why that breaks traditional approaches and how semantic reading solves it.
Why Daangn Market Screenshots Break Traditional OCR
Hangul in item names requires Korean-capable character recognition, not just English OCR. Listing titles like "삼성 냉장고 2022" or "에어팟 프로 2세대" mix Korean syllables, English brand names, and numerals in the same text block. Standard OCR engines trained primarily on Latin scripts produce garbled output for the Korean portions, breaking the Item Name field entirely.
The currency format uses ₩ symbols, commas for thousands, and optional '원' suffix. Daangn displays prices as "₩1,200,000" in some UI versions and "1,200,000원" in others — both use Korean thousand separators (commas) rather than Western period-separated formats. A tool expecting "$1,200" or "£65.00" may misread or drop the Korean price entirely.
Location uses a hyperlocal Korean administrative hierarchy that most extractors do not understand. Daangn shows town/neighborhood level — "서울시 마포구" or "성남시 분당구 정자동" — rather than a simple city name. Traditional OCR reports this as an unstructured text blob without recognizing it as a location field, mixing it with nearby UI labels.
How Semantic Extraction Reads Korean Listing Card Layouts
Name the fields — the AI reads Hangul titles by visual role, not by character set. Type Item Name, Price, Location. The visual model identifies the boldest card text — whether it is "아이폰 15 프로 팝니다" or "맥북에어 M1 2020" — as the Item Name by its size and position in the card layout, and reads it with Korean-capable recognition.
One column definition works across ₩ and '원' price variants. Upload a mix of Daangn screenshots — some showing "₩12,000" and others "12,000원" — into the same batch. The AI reads each independently, recognizing the currency indicator by its context in the card (adjacent to the largest numeric block in a specific zone), and outputs the raw display value in the Price column.
Process gallery screenshots alongside fresh Daangn captures. A screenshot from last week's Daangn search still works even if the listing has been sold or removed. Mix listing cards, detail pages, and chat preview screenshots in one batch — the AI adapts to each layout independently and outputs every row with the fields you defined.
From a Gallery of Daangn Market Screenshots to One Consolidated Spreadsheet
Whether you are tracking prices across Seoul neighborhoods, archiving completed transactions, or compiling market research data, here is how the recognition workflow turns your screenshot collection into a structured record.
Upload Your Daangn Market Screenshots
Drop in the captures from your gallery — individual listing cards from the Daangn app, search results showing multiple items, or detail pages with full descriptions. PNG, JPG, or WebP from any device. Mix screenshots from the Korean 당근마켓 app and the international Karrot version in the same batch — no sorting by app variant needed beforehand.
Define the Fields You Need
Type the columns for your tracking spreadsheet: Item Name, Price, Location. The AI reads each screenshot independently — a listing from Seoul yields "아이폰 15 프로" with "₩1,200,000" and "서울시 마포구"; a listing from Seongnam yields "맥북에어 M1" with "₩850,000" and "성남시 분당구". Fields that do not appear on a given screenshot remain blank.
Export a Consolidated Listing 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 the columns you defined. Filter by Location to compare prices across different Seoul districts. Sort by Price to find the best deals across neighborhoods. Export directly to Excel or push to Google Sheets via the add-on for ongoing tracking.
When Daangn Market Screenshot Recognition Works Best — and When to Be Cautious
When It Works Best
Full-resolution native screenshots of listing cards or detail pages. Machine-rendered Hangul and Korean text from the Daangn app reaches up to 99% accuracy for fields like Item Name and Price.
Mixed Daangn and Karrot branding variants in one batch. Screenshots from the Korean 당근마켓 app and the international Karrot app processed together — the AI adapts to each card style independently.
Gallery screenshots of listings no longer live in the app. Since the AI reads the visual output, a saved screenshot captures data even after the item is sold or the listing is removed from Daangn Market.
When to Be Cautious
Screenshots forwarded through messaging apps. KakaoTalk, LINE, and WhatsApp compress images, reducing accuracy for smaller Korean text characters. Use native screenshots for best results.
Multiple listings in one screenshot. The AI reads the most visually prominent card. For clean per-item extraction, capture each listing card individually.
Photos of a screen taken at an angle. Glare and distortion reduce accuracy for both Hangul and numeric fields. Capture native screenshots on your phone for best results.
Frequently Asked Questions
Can the tool recognize Korean (Hangul) item names from Daangn Market screenshots accurately?
Yes. The AI is trained to read Korean Hangul and mixed Korean-English text commonly found in Daangn listing titles. Item names like 아이폰 15 프로 팝니다 or 맥북에어 M1 2020 are recognized at high accuracy. The Item Name field captures the title exactly as displayed in the screenshot. For best results, use native full-resolution screenshots rather than images compressed through messaging apps such as KakaoTalk or LINE.
How does the tool handle the ₩ currency format used in Daangn Market — for example, 12,000원?
The AI reads the full Korean price display including the ₩ symbol or the 원 suffix and the comma-separated thousands format. A listing showing 12,000원 or ₩12,000 is captured in the Price column exactly as rendered on the screenshot — including the currency indicator. The raw display text is preserved so you can strip symbols or convert units in your spreadsheet after export. Both formatting variants are handled in the same batch without any configuration changes.
Does it work with both "Daangn Market" and "Karrot" branded screenshots — the app uses different names in Korea versus international markets?
Yes. The app is branded 당근마켓 (Daangn Market) in South Korea and Karrot in international markets such as the UK, Canada, and Japan. The underlying listing card structure — item photo, title, price, location row — is consistent across both branding variants. The AI reads the listing card data, not the app header or logo, so the same column definitions work regardless of whether the screenshot comes from the Korean Daangn version or the international Karrot version.
Does the Location field extract neighborhood-level detail like 서울시 마포구 rather than just the city name?
Yes. Daangn Market operates on a hyperlocal model — each listing shows the town or neighborhood level (동/구) such as 서울시 마포구 or 성남시 분당구 rather than just a broad city name. The AI captures this hyperlocal location string exactly as displayed on the listing card. If the screenshot includes the full detail page, the location may also include the specific dong (neighborhood) name — for example, 분당구 정자동 — for even more granularity. Daangn never shows the exact street address, only this district-level location, which is consistent with how the app shares location for privacy reasons.
Can I process gallery screenshots taken weeks ago alongside fresh Daangn Market captures in the same batch?
Yes. Since the AI reads the visual content of the screenshot rather than querying a live API or scraping the app, a Daangn Market listing screenshot captured last month still works even if the item has been sold or removed from the app. Mix old gallery screenshots with fresh captures in the same batch — each image is processed independently, and every row in the output spreadsheet corresponds to one screenshot. This makes it easy to consolidate data from screenshots accumulated over time into a single structured file.
Return to the secondhand marketplace overview — covers Facebook Marketplace, Vinted, Carousell, OfferUp, and more.
Deep dives: Batch processing marketplace screenshots into a single spreadsheet · General guide to extracting data from screenshots