Ebay Kleinanzeigen API

<div align="center"> <h3 align="center">Ebay Kleinanzeigen API</h3> <p align="center"> A powerful API interface for Ebay-Kleinanzeigen.de that enables you to fetch listings and specific data. </p> <p align="center"> <b>🚀 Looking for a ready-to-use solution?</b> <br /> Try it at <a href="https://kleinanzeigen-agent.de/features/developer-api"><strong>kleinanzeigen-agent.de »</strong></a> <br /> ✓ Automated Search Agents <br /> ✓ Search & Detail API <br /> <a href="https://github.com/DanielWTE/ebay-kleinanzeigen-api/issues">Report Bug</a> · <a href="https://github.com/DanielWTE/ebay-kleinanzeigen-api/issues">Request Feature</a> </p> </div> <p align="right">(<a href="#readme-top">back to top</a>)</p>

Getting Started

Want to skip the setup?

Visit kleinanzeigen-agent.de for our hosted solution with additional features and zero configuration required.

Prerequisites

Python 3.12 or higher
pip (Python package manager)
Playwright

Installation

Clone the repository

sh
git clone https://github.com/DanielWTE/ebay-kleinanzeigen-api.git
cd ebay-kleinanzeigen-api

Install dependencies

sh
pip install -r requirements.txt
playwright install chromium

Start the API

sh
uvicorn main:app --reload

The API will be available at http://localhost:8000

Non-headless mode (visible browser)

Set the HEADLESS environment variable to false to watch the browser while scraping — useful during development:

sh
HEADLESS=false uvicorn main:app --reload

Installation using UV

Prerequisites

uv package manager

Installation of dependencies

sh
uv sync

Start the API

sh
uv run uvicorn main:app --reload

Non-headless mode with UV

sh
HEADLESS=false uv run uvicorn main:app --reload

Installation using Docker

Build the Docker image

sh
docker build -t ebay-kleinanzeigen-api .

Run the Docker container

sh
docker run -p 8000:8000 ebay-kleinanzeigen-api

The API will be available at http://localhost:8000

API Endpoints

1. Fetch Listings (Standard)

Endpoint: GET /inserate

Description: Retrieves a list of listings based on search criteria.

Query Parameters

query (string, optional): The search term (e.g., "fahrrad" to search for bicycles).
location (string, optional): The location or postal code to narrow the search (e.g., 10178 for Berlin).
radius (integer, optional): The search radius in kilometers from the specified location (e.g., 5 for a 5 km radius).
min_price (integer, optional): The minimum price in Euros for the listings (e.g., 200 for at least 200 Euros).
max_price (integer, optional): The maximum price in Euros for the listings (e.g., 500 for at most 500 Euros).
page_count (integer, optional): The number of pages to search or return (e.g., 5 for the first 5 pages, default is 1, max: 20 pages).
min_publish_date (datetime, optional): Stop fetching once a page contains listings published before this datetime. Listings older than the threshold are removed from the final results. Format: YYYY-MM-DDTHH:MM:SS (e.g., 2026-05-03T08:00:00). Useful for intraday runs that should only collect new listings.

Example Requests

http
GET /inserate?query=fahrrad&location=10178&radius=5&min_price=200&page_count=5

With min_publish_date — stop when listings older than this morning are reached:

sh
curl "http://localhost:8000/inserate?query=fahrrad&page_count=10&min_publish_date=2026-05-04T08:00:00"

2. Fetch Listing Details

Endpoint: GET /inserat/{id}

Description: Retrieves detailed information about a specific listing.

Path Parameters

id (string): The listing identifier. Two formats are accepted:

Format Example Source
Full URL segment (recommended) 3399373623-220-16792 Last path segment of the listing url field from search results
Plain adid 3399373623 The adid field from search results

Kleinanzeigen listing URLs follow the pattern:
```
/s-anzeige/{slug}/{adid}-{category_id}-{location_id}
```
The suffix -{category_id}-{location_id} is used by Kleinanzeigen for server-side routing (CDN/load balancer can route directly without a database lookup). Using the full segment avoids a potential server-side redirect — Kleinanzeigen may or may not redirect depending on the listing.

Format	Example	Source
Full URL segment (recommended)	`3399373623-220-16792`	Last path segment of the listing `url` field from search results
Plain adid	`3399373623`	The `adid` field from search results

curl Examples

With full URL segment (recommended, ~4s):

sh
curl -s http://localhost:8000/inserat/3399373623-220-16792 | python -m json.tool

With plain adid only (also works, redirect may or may not occur):

sh
curl -s http://localhost:8000/inserat/3399373623 | python -m json.tool

Tip: When iterating over search results, use the url field directly — extract the last path segment to get the full id:

sh
# Extract the id from a listing url
echo "https://www.kleinanzeigen.de/s-anzeige/fendt-bianco/3399373623-220-16792" | grep -oP '[^/]+$'
# → 3399373623-220-16792

Example Response

json
{
  "success": true,
  "time_taken": 4.108,
  "data": {
    "id": "3399373623",
    "url_requested": "https://www.kleinanzeigen.de/s-anzeige/3399373623-220-16792",
    "url_redirected": "https://www.kleinanzeigen.de/s-anzeige/fendt-bianco-aktiv-550-kg/3399373623-220-16792",
    "categories": ["Wohnwagen & Reisemobile", "Wohnwagen"],
    "title": "Fendt Bianco Aktiv 550 KG",
    "status": "active",
    "price": {
      "amount": "12500",
      "currency": "€",
      "negotiable": true
    },
    "delivery": "pickup",
    "location": {
      "zip": "88299",
      "city": "Leutkirch im Allgäu"
    },
    "views": "142",
    "description": "Verkaufe unseren gepflegten Fendt Bianco...",
    "images": [
      "https://img.kleinanzeigen.de/api/v1/prod-ads/images/..."
    ],
    "details": {
      "Baujahr": "2015",
      "Zustand": "Gebraucht"
    },
    "features": {},
    "seller": {
      "name": "Max Mustermann",
      "type": "private"
    },
    "extra_info": {}
  },
  "performance_metrics": {
    "success_rate": 100,
    "time_taken": 4.108
  }
}

Response fields:

data.id — Kleinanzeigen listing ID (numeric string).
data.url_requested — The URL sent to Kleinanzeigen. When called with a plain adid, this is the short form without slug.
data.url_redirected — The final URL after any server-side redirect. When called with a plain adid, this contains the full canonical URL including slug and routing suffix {adid}-{category_id}-{location_id}. Identical to url_requested when the full segment was passed.
data.status — Listing status: active, sold, reserved, or deleted.
data.price.negotiable — true if the price is marked as negotiable (VB).
data.delivery — pickup, shipping, or null.
data.details — Key/value pairs of listing attributes (e.g. Baujahr, Zustand).
data.features — Additional configuration options if present.

3. Fetch Listings with Details (Combined)

Endpoint: GET /inserate-detailed

Description: Retrieves listings and their detailed information in a single request.

Query Parameters

Same as /inserate endpoint, plus:

max_concurrent_details (integer, optional): Maximum concurrent detail fetches (default: 5, max: 10).

Example Request

http
GET /inserate-detailed?query=laptop&page_count=2&max_concurrent_details=5

4. Scrape by URL

Endpoint: POST /inserate-by-url

Description: Scrape listings using a full Kleinanzeigen search URL. All filters already encoded in the URL (category, brand, year, fuel type, transmission, car type, etc.) are preserved as-is. Page numbers are injected automatically for multi-page fetching. Use this when /inserate does not yet support a filter you need.

Pages are fetched sequentially to avoid IP-level rate limiting by Kleinanzeigen.

Request Body

json
{
  "url": "https://www.kleinanzeigen.de/s-autos/volkswagen/klima/k0c216+autos.marke_s:volkswagen",
  "max_pages": 3
}

url (string, required): Any Kleinanzeigen search or category URL.
max_pages (integer, optional): Number of pages to fetch (default: 1). Each page returns up to 25 listings.
min_publish_date (datetime, optional): Stop fetching once a page contains listings published before this datetime. Listings older than the threshold are removed from the final results. Format: YYYY-MM-DDTHH:MM:SS. Supports intraday precision — e.g. "2026-05-04T08:00:00" keeps only listings from that morning onward.

curl Examples

Single page — VW with Klima (25 results):

sh
curl -X POST http://localhost:8000/inserate-by-url \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://www.kleinanzeigen.de/s-autos/volkswagen/klima/k0c216+autos.marke_s:volkswagen",
    "max_pages": 1
  }'

3 pages — VW with Klima (up to 75 results):

sh
curl -X POST http://localhost:8000/inserate-by-url \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://www.kleinanzeigen.de/s-autos/volkswagen/klima/k0c216+autos.marke_s:volkswagen",
    "max_pages": 3
  }'

With full filters — VW Kombi/SUV, Automatik, CNG/LPG, ab 2008, 3 pages:

sh
curl -X POST http://localhost:8000/inserate-by-url \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://www.kleinanzeigen.de/s-autos/volkswagen/klima/k0c216+autos.ez_i:2008%2C+autos.fuel_s:(cng%2Clpg)+autos.marke_s:volkswagen+autos.shift_s:automatik+autos.typ_s:(kombi%2Csuv)",
    "max_pages": 3
  }'

Wohnwagen — Fendt, Klima, max 15.000 €, ab 2008:

sh
curl -X POST http://localhost:8000/inserate-by-url \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://www.kleinanzeigen.de/s-wohnwagen-mobile/wohnwagen/preis::15000/klima/k0c220+wohnwagen_mobile.art_s:wohnwagen+wohnwagen_mobile.ez_i:2008%2C+wohnwagen_mobile.marke_s:fendt",
    "max_pages": 2
  }'

With min_publish_date — stop once listings older than this morning are found (intraday run):

sh
curl -s -X POST http://localhost:8000/inserate-by-url \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://www.kleinanzeigen.de/s-autos/volkswagen/klima/k0c216+autos.marke_s:volkswagen",
    "max_pages": 20,
    "min_publish_date": "2026-05-04T08:00:00"
  }' | python -m json.tool

Expected: pages_requested < 20, all published_at values ≥ "2026-05-04T08:00:00".

Verify early-stop: future cutoff → 0 results, only 1 page fetched:

sh
curl -s -X POST http://localhost:8000/inserate-by-url \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://www.kleinanzeigen.de/s-autos/volkswagen/klima/k0c216+autos.marke_s:volkswagen",
    "max_pages": 5,
    "min_publish_date": "2099-01-01T00:00:00"
  }' | python -m json.tool

Expected: "unique_results": 0, "pages_requested": 1.

Verify no false filtering: past cutoff → full 25 results:

sh
curl -s -X POST http://localhost:8000/inserate-by-url \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://www.kleinanzeigen.de/s-autos/volkswagen/klima/k0c216+autos.marke_s:volkswagen",
    "max_pages": 1,
    "min_publish_date": "2000-01-01T00:00:00"
  }' | python -m json.tool

Expected: "unique_results": 25.

Verify schema enforcement: plain date string → HTTP 422:

sh
curl -s -X POST http://localhost:8000/inserate-by-url \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://www.kleinanzeigen.de/s-autos/volkswagen/klima/k0c216+autos.marke_s:volkswagen",
    "max_pages": 1,
    "min_publish_date": "2026-05-04"
  }' | python -m json.tool

Expected: HTTP 422 — the time component is required.

Pretty-print the response with Python:

sh
curl -s -X POST http://localhost:8000/inserate-by-url \
  -H "Content-Type: application/json" \
  -d '{"url": "https://www.kleinanzeigen.de/s-autos/volkswagen/klima/k0c216+autos.marke_s:volkswagen", "max_pages": 1}' \
  | python -m json.tool

Example Response

json
{
  "success": true,
  "results": [
    {
      "adid": "3398033022",
      "url": "https://www.kleinanzeigen.de/s-anzeige/...",
      "title": "VW Golf Variant 2.0 TDI Automatik Klima",
      "price": "12500",
      "description": "Gepflegter Zustand, Erstbesitz...",
      "published_at": "2026-05-03T22:06:00"
    }
  ],
  "unique_results": 75,
  "total_results": 120140,
  "time_taken": 12.358,
  "performance_metrics": {
    "pages_requested": 3,
    "pages_successful": 3,
    "success_rate": 100.0,
    "average_page_time": 4.119
  },
  "browser_metrics": {
    "contexts_created": 3,
    "contexts_reused": 0,
    "contexts_in_pool": 3,
    "contexts_in_use": 0
  }
}

Response fields:

results: Array of listing objects. Each listing contains:
- adid — Kleinanzeigen internal listing ID
- url — Direct link to the listing
- title — Listing title
- price — Price as string (numeric digits only, no currency symbol; empty string if not set)
- description — Short teaser text from the listing card
- published_at — ISO 8601 publish datetime (e.g. "2026-05-03T22:06:00"). Older listings show only the date at midnight ("2026-04-26T00:00:00"). null if not shown on the card.
unique_results: Total number of listings returned across all fetched pages.
total_results: Total hits reported by Kleinanzeigen for this search (e.g. 120140). Extracted from the page breadcrumb on the first page load — no extra request needed.
time_taken: Total elapsed time in seconds.
performance_metrics.pages_requested / pages_successful / success_rate: Page-level fetch statistics.

5. Convert URL to API Parameters

Endpoint: POST /convert-url

Description: Parses a Kleinanzeigen URL and returns two groups of parameters: inserate_params (what the /inserate endpoint currently understands) and unmapped (filters not yet supported by /inserate, such as category, brand, year, or transmission). Useful for inspecting which parts of a URL can be expressed via the structured API.

Request Body

json
{
  "url": "https://www.kleinanzeigen.de/s-autos/volkswagen/klima/k0c216+autos.marke_s:volkswagen"
}

curl Examples

Basic — category + keyword only:

sh
curl -X POST http://localhost:8000/convert-url \
  -H "Content-Type: application/json" \
  -d '{"url": "https://www.kleinanzeigen.de/s-autos/volkswagen/klima/k0c216+autos.marke_s:volkswagen"}' \
  | python -m json.tool

With year filter:

sh
curl -X POST http://localhost:8000/convert-url \
  -H "Content-Type: application/json" \
  -d '{"url": "https://www.kleinanzeigen.de/s-autos/volkswagen/klima/k0c216+autos.ez_i:2008%2C+autos.marke_s:volkswagen"}' \
  | python -m json.tool

With full filters — fuel, transmission, body type:

sh
curl -X POST http://localhost:8000/convert-url \
  -H "Content-Type: application/json" \
  -d '{"url": "https://www.kleinanzeigen.de/s-autos/volkswagen/klima/k0c216+autos.ez_i:2008%2C+autos.fuel_s:(cng%2Clpg)+autos.marke_s:volkswagen+autos.shift_s:automatik+autos.typ_s:(kombi%2Csuv)"}' \
  | python -m json.tool

Example Response

json
{
  "inserate_params": {
    "query": "klima",
    "page_count": 1
  },
  "unmapped": {
    "category_slug": "s-autos",
    "subcategory": "volkswagen",
    "category_id": 216,
    "year_from": 2008,
    "year_to": null,
    "brands": ["volkswagen"]
  }
}

Documentation

API Response Format

`GET /inserate` and `GET /inserate-detailed`

json
{
  "success": true,
  "time_taken": 1.23,
  "unique_results": 25,
  "data": [
    {
      "adid": "123456",
      "url": "https://www.kleinanzeigen.de/s-anzeige/...",
      "title": "Example Item",
      "price": "100",
      "description": "Short teaser text..."
    }
  ],
  "performance_metrics": {
    "pages_requested": 1,
    "pages_successful": 1,
    "success_rate": 100.0,
    "average_page_time": 1.23
  }
}

The /inserate-detailed response additionally nests a details object inside each listing with full description, seller info, and location.

`POST /inserate-by-url`

json
{
  "success": true,
  "results": [
    {
      "adid": "123456",
      "url": "https://www.kleinanzeigen.de/s-anzeige/...",
      "title": "Example Item",
      "price": "100",
      "description": "Short teaser text...",
      "published_at": "2026-05-03T22:06:00"
    }
  ],
  "unique_results": 75,
  "total_results": 120140,
  "time_taken": 12.358,
  "performance_metrics": {
    "pages_requested": 3,
    "pages_successful": 3,
    "success_rate": 100.0,
    "average_page_time": 4.119
  },
  "browser_metrics": { ... }
}

total_results is the total hit count shown by Kleinanzeigen (e.g. 120140). It is extracted from the breadcrumb of the first page — no extra request is made. The field is omitted if the breadcrumb is not found.

`POST /convert-url`

json
{
  "inserate_params": {
    "query": "klima",
    "page_count": 1
  },
  "unmapped": {
    "category_slug": "s-autos",
    "subcategory": "volkswagen",
    "category_id": 216,
    "year_from": 2008,
    "year_to": null,
    "brands": ["volkswagen"]
  }
}

Performance Features

📊 Performance Benchmarks

Performance tests conducted on Arch Linux with Intel i7-1260P processor:

Endpoint	Operation	Avg Time	Min Time	Max Time	Success Rate	Results
Root	API status check	0.004s	0.001s	0.009s	100%	-
Listings	1 page search	1.209s	1.081s	1.596s	100%	25
Listings	5 pages search	2.436s	2.328s	2.532s	100%	125
Listings	10 pages search	4.989s	4.790s	5.229s	100%	250
Details	Single listing details	1.072s	0.928s	1.276s	100%	1
Combined	1 page + details	9.058s	8.804s	9.319s	100%	25
Combined	2 pages + details	17.512s	17.063s	18.038s	100%	50

Performance Highlights

Single page search: ~1.2s average response time
10-page search: ~5.0s average response time
Individual listing details: ~1.1s average response time
Combined search + details: ~9.1s average for 1 page with full details

All tests performed with uvloop optimization and browser context pooling enabled.

🚀 Technical Features

uvloop Integration: High-performance asyncio event loop
Context Pooling: Efficient browser resource reuse
Memory Optimization: Automatic garbage collection and efficient processing
Intelligent Concurrency: Optimal concurrent processing with resource control

🔧 Production Ready

Clean API Responses: Minimal overhead with essential data only
Resource Management: Automatic cleanup and efficient resource usage
Scalable Architecture: Handles concurrent requests efficiently
Docker Optimized: Fast container builds with uv package manager

API documentation is available at http://localhost:8000/docs when running locally.

License

Distributed under the MIT License. See LICENSE for more information.

Changelog

May 2026 — Multi-image fetching, deleted-ad detection, media response restructure ⚠️ Breaking

Maintainer: Darko Palić

⚠️ Breaking changes

1. `GET /inserat/{id}` — response shape changed

The images and views top-level fields have been replaced. The data is still present
but lives in a different location.

Field	Before	After
Image URLs	`data.images` → `["url"]` (one image)	`data.media.images.urls` → `["url1", "url2", …]` (all images)
Image count	not provided	`data.media.images.count`
Views	`data.views`	`data.extra_info.views` (unchanged location — was already there)

How to migrate:

python
# Before
image = response["data"]["images"]        # single URL string or list of one
views = response["data"]["views"]

# After
images = response["data"]["media"]["images"]["urls"]   # list of all gallery images
count  = response["data"]["media"]["images"]["count"]
views  = response["data"]["extra_info"]["views"]       # unchanged

javascript
// Before
const image = data.images;
const views = data.views;

// After
const images = data.media.images.urls;   // array of all gallery images
const count  = data.media.images.count;
const views  = data.extra_info.views;    // unchanged

2. `GET /inserat/{id}` — deleted/expired ads now return HTTP 404

Previously, fetching a deleted or expired ad returned HTTP 200 with success: true
and empty fields (title "[ERROR] Title not found", status "active").

Now, two conditions are detected and return HTTP 404 with success: false:

The URL redirects away from the ad page (ad deleted — Kleinanzeigen redirects to homepage)
The page contains #srchrslt-adexpired (ad expired page)

How to migrate:

python
# Before — callers had to inspect fields to detect a missing ad
resp = requests.get(f"/inserat/{id}", params={"batch_id": "x"})
data = resp.json()["data"]
if data["title"] == "[ERROR] Title not found":
    # ad was gone

# After — standard HTTP semantics
resp = requests.get(f"/inserat/{id}", params={"batch_id": "x"})
if resp.status_code == 404:
    # ad is deleted or expired
    print(resp.json()["detail"]["status"])  # "deleted"
else:
    data = resp.json()["data"]

javascript
// Before
const res = await fetch(`/inserat/${id}?batch_id=x`);
const { data } = await res.json();
if (data.title === "[ERROR] Title not found") { /* gone */ }

// After
const res = await fetch(`/inserat/${id}?batch_id=x`);
if (res.status === 404) {
    const { detail } = await res.json();
    // detail.status === "deleted"
} else {
    const { data } = await res.json();
}

New fields (non-breaking, additive)

data.scraped_at — ISO 8601 UTC timestamp of when the ad was scraped
data.url_requested — URL sent to Kleinanzeigen (built from the id)
data.url_redirected — final URL after any server-side redirect
data.seller.user_id — numeric Kleinanzeigen user ID extracted from the profile link href

Improvements (non-breaking)

All gallery images returned — previously only the first image was fetched
(#viewad-image). Now all images from .galleryimage-element img are returned,
including data-src lazy-loaded images. Deduplication applied.
POST /inserate-by-url — smarter pagination — breadcrumb is now fully parsed
on page 1 to determine the exact page count. Fetching stops automatically when the
real page count is reached, avoiding wasted requests for non-existent pages.
POST /inserate-by-url — URL pagination fix — page injection now correctly
handles complex URLs with filters before the category segment
(e.g. anzeige:angebote, preis::N).
Healthcheck endpoint — GET /health added for load balancer / uptime monitoring.

Tests

16 live integration tests in tests/test_inserat_live.py covering active ads
(both full-segment and plain adid formats), all response fields, and deleted-ad
detection (HTTP 404, status: "deleted") for 2 known-expired ad IDs.

May 2026 — `GET /inserat/{id}` URL tracking (`url_requested` / `url_redirected`)

Maintainer: Darko Palić

Changed

GET /inserat/{id} response now includes two URL fields in data:
- url_requested — the URL that was sent to Kleinanzeigen (built from the id passed to the endpoint).
- url_redirected — the final URL as reported by the browser after any server-side redirect. Identical to url_requested when no redirect occurred.
Useful when calling with a plain adid (e.g. /inserat/3399373623) to discover the canonical URL including slug and routing suffix.

Tests

13 live integration tests in tests/test_inserat_live.py covering both id formats (full segment and plain adid), url_requested, url_redirected, content fields, and status validation.

May 2026 — Intraday `min_publish_date` precision (datetime)

Maintainer: Darko Palić

Changed

min_publish_date upgraded from date to datetime on both GET /inserate and POST /inserate-by-url. The parameter now accepts a full ISO 8601 datetime string (YYYY-MM-DDTHH:MM:SS) instead of a plain date. This enables intraday filtering — e.g. "2026-05-04T08:00:00" keeps only listings published from that exact time onward, not just from that calendar day.
The internal comparison in _page_has_old_listings and _filter_by_min_publish_date now operates on full datetime objects instead of stripping to .date(), preserving hour/minute precision.
German date strings ("Heute, 08:46", "Gestern, 14:29") already produced full ISO 8601 datetimes — no parsing changes required.

Tests

3 new live integration tests in tests/test_inserate_by_url_live.py covering min_publish_date:
- Future cutoff (2099-01-01T00:00:00) returns 0 results and stops after page 1.
- Past cutoff (2000-01-01T00:00:00) does not filter anything — result count unchanged.
- All returned published_at values satisfy the cutoff when a real intraday threshold is used.

May 2026 — URL-passthrough scraping, publish date, rate-limit fixes

Maintainer: Darko Palić

New endpoints

POST /inserate-by-url — scrape listings using any full Kleinanzeigen search URL. All filters encoded in the URL (category, brand, year, fuel type, transmission, body type, etc.) are preserved as-is. Page numbers are injected automatically for multi-page fetching. Response includes total_results (total hit count from the page breadcrumb, no extra request) and published_at on every listing.
POST /convert-url — parse a Kleinanzeigen URL into inserate_params (what /inserate understands today) and unmapped (filters not yet supported by the structured endpoint).

New listing field

published_at — ISO 8601 publish datetime added to every listing result (e.g. "2026-05-04T22:06:00"). Extracted from the listing card during the existing page load — no extra request. Older listings return a date-only value ("2026-04-26T00:00:00"); null if not shown on the card.

Bug fixes

Cookie clearing on context release — browser contexts now have session cookies cleared before returning to the pool, preventing Kleinanzeigen tracking state from carrying over between unrelated requests and causing silent empty results on pages 2+.
Inter-page delay — a 2-second pause between sequential page fetches within a single multi-page scrape reduces bot-detection false positives.

Tests

11 unit tests for /convert-url (tests/test_convert_url.py, no server required)
14 live integration tests for /inserate-by-url (tests/test_inserate_by_url_live.py), covering JSON structure, published_at, total_results, and exact result counts for 1–4 pages (25 / 50 / 75 / 100 results)