Portál Dražeb Auction Scraper

Extract Czech auctions from Portál Dražeb (portaldrazeb.cz) into structured JSON, CSV or Excel data. Find upcoming property auctions, land, apartments, vehicles and other assets with starting bids, auction dates, categories, counties and source links.

Start with Auction status and Results limit. The default returns one upcoming auction. No Portál Dražeb API token or AI service is required.

What you can do

• Build a searchable export of public Czech auctions.
• Filter by asset category, county, start date and starting bid in CZK.
• Add full auction descriptions, appraisal values, deposits, available images and public document links.
• Extract text from selected PDF documents without an AI surcharge.
• Compare repeated runs and return only new or changed auctions.

Quick start

1. Choose an auction status, such as Upcoming.
2. Set your results limit and optionally add filters.
3. Click Start, then download the dataset from the Output tab.

{
  "sourceViews": ["upcoming"],
  "maxAuctions": 10
}

The fast default collects listing information. Turn on Include auction details for additional fields and document links. Selecting Extract PDF text enables detail collection automatically.

Inputs

Setting	What it does
Auction status	Upcoming, in progress, ended, online, or overbids. Multiple views share one result limit.
Results limit	Maximum auctions delivered, from 1 to 1,000. Default: 1.
Asset categories	Property, land, houses, apartments, vehicles and other source categories.
Asset / auction counties	Czech county names as displayed by the source, such as Třebíč.
Starts from / until	Inclusive Prague calendar dates (YYYY-MM-DD) or timestamps with a timezone.
Minimum / maximum starting price	CZK bounds. Auctions with unknown starting bids do not match a price filter.
Sort by start time	Earliest or latest first within each selected view.
Include auction details	Visit detail pages for available additional fields and public document links. Default: off.
PDF content	Links only, or extract PDF text. Caps default to 2 PDFs per auction and 10 per run.
Results to return	All current auctions, new only, or new and changed. Default: all.
Monitoring key	Optional name for your repeated comparison stream.
Maximum listing pages	Global safety limit, default 50, maximum 200.

Category and county filters use the source's public form. Date and price boundaries are checked against displayed listing values. Collection stops at the result limit, customer charge limit, listing-page limit, source end or run time budget. Tight filters can return no matches within the inspected pages; check the run summary for coverage.

Example with filters and optional documents:

{
  "sourceViews": ["upcoming"],
  "categories": ["land"],
  "itemCounties": ["Třebíč"],
  "maxAuctions": 5,
  "maxPriceCzk": 2000000,
  "includeDetails": true,
  "documentMode": "text",
  "maxPdfDocumentsPerAuction": 1,
  "maxPdfDocumentsTotal": 3
}

Output

Each auction has recordType: "auction". When details contain public attachments, linked document rows have recordType: "document". Document rows are included in the auction price.

A shortened auction example:

{
  "recordType": "auction",
  "auctionId": "233ex200-25-35-d5drd",
  "auctionNumber": "233EX200/25-35",
  "title": "stavební pozemek v obci Jaroměřice nad Rokytnou, okres Třebíč",
  "categoryTitleCzech": "Pozemek",
  "startingPriceCzk": 486667,
  "startsAtLocalText": "29.10.2026 10:00:00",
  "startsAt": "2026-10-29T09:00:00.000Z",
  "assetCountyCzech": "Třebíč",
  "detailStatus": "notRequested",
  "sourceUrl": "https://www.portaldrazeb.cz/drazba/233ex200-25-35-d5drd",
  "dataInputSource": "web",
  "sourceVerified": true
}

Detail mode adds available fields such as appraisedValueCzk, depositCzk, descriptionCzech, documentUrls and imageUrls. Unavailable values remain null. Dates retain the original Czech text; UTC conversion uses Europe/Prague and leaves ambiguous daylight-saving times unknown.

Document rows include auctionId, url, originalName, extractionStatus and extractedTextCzech. PDF extraction is limited to 15 MB per document and 200,000 text characters. Image-only scans are not OCRed. A failed document retains its link and status without discarding the auction.

The Run summary reports delivered counts, source coverage, stop reason, detail/PDF failures, monitoring and billing. A literal phrase in Find a phrase in results produces matching auction links in Auction results analysis, using only delivered results.

New and changed auctions

Use a consistent monitoring key and filters across repeated runs:

{
  "sourceViews": ["upcoming"],
  "maxAuctions": 20,
  "outputMode": "newAndChanged",
  "monitoringKey": "my-property-watch"
}

The first run establishes a baseline from delivered results. Subsequent runs classify auctions as new, changed or unchanged, with changed field groups. Comparison streams are isolated by account, filters and detail profile. Changing those settings starts a separate baseline. Unchanged auctions are not charged as auction results when suppressed. Previously observed records expire after 180 days without observation. Capped searches do not imply auctions were removed.

Import existing auctions

The optional Import existing auctions section accepts auction objects in providedRecords or an accessible Apify dataset ID in providedDatasetId. Include a title and source URL; use the dataset field names for any additional values. Set JSON records to [] when using a dataset. At most 1,000 input records are read.

Nonempty supplied records replace live collection. The Actor validates and deduplicates them, applies the requested output/monitoring mode and marks sourceVerified: false. Imports do not visit source websites or download PDFs. Delivered imported auctions use the same auction-result price.

Pricing and run size

$2 per 1,000 delivered auctions, plus $0.001 per run, subject to the current Pricing tab and your plan. Linked document rows, local PDF text extraction and change classification have no separate event fee. Set Maximum cost per run to cap spending; the Actor stops collecting or enriching when no further auction results fit.

The default uses 1 GB memory to give the public browser sufficient headroom. Live collection requires at least 512 MB. Small listing exports are faster than detail and PDF runs. Large searches can stop before the requested count when safety or source limits are reached.

Use the API with Python

import os
from apify_client import ApifyClient

client = ApifyClient(os.environ["APIFY_TOKEN"])
run = client.actor("thescrapelab/portal-drazeb-auction-scraper").call(
    run_input={"sourceViews": ["upcoming"], "maxAuctions": 10},
    memory_mbytes=1024,
    timeout_secs=300,
    max_total_charge_usd=0.03,
)
if run is None or run["status"] != "SUCCEEDED":
    raise RuntimeError("Auction collection did not succeed; inspect the run log.")
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    if item["recordType"] == "auction":
        print(item["title"], item["startingPriceCzk"], item["sourceUrl"])

Coverage and support

The Actor reads public rendered listings and detail pages. It does not log in, place bids or retrieve private bidder/payment information. Descriptions remain in Czech. Source availability, blocking, layout changes and unavailable attachments can affect coverage. It collects selected views sequentially and deduplicates auctions across them; sorting is per view rather than a combined global ranking.

For investment or bidding decisions, verify the linked source and auction decree. Use the Store Issues tab with a run link and the expected result if you need help.