What's New in ScrapeUnblocker: September 2026

This is the first monthly roundup on the ScrapeUnblocker blog. Once a month we collect what shipped publicly, group it by theme, and focus on the part that matters to you: what you can build with it.

This one covers late August to the end of September, and it was a busy stretch. Most of the work went into dedicated plugins. A plugin takes a few plain inputs, like a keyword, a city or a flight route, and returns structured JSON. You never parse HTML and you never deal with the anti-bot layer yourself.

September at a Glance

• Retail: product search plugins for Walmart, Target, Best Buy, The Home Depot, Etsy, SHEIN, AliExpress and Zalando.
• Marketplaces: Vinted, Craigslist and leboncoin listings as JSON.
• Travel: Southwest Airlines fares in dollars or points, Booking.com hotels and Airbnb stays.
• Jobs: Indeed, Naukri and LinkedIn postings with parsed salaries and filters.
• Property and cars: Zillow, Redfin, Rightmove, ImmobilienScout24, mobile.de and AutoScout24.
• Social, video and ads: a five-endpoint TikTok plugin, YouTube search and comments, Threads profiles and the Meta Ad Library.
• Local, apps and reviews: Google Maps place details, Yelp, Trustpilot, Google Play and Product Hunt.
• Platform: bigger Google Images results, browser steps, a ChatGPT connector and clearer error codes.

Every plugin below has a page in the scraper catalogue with its fields, a sample request and its public benchmark. You call all of them with your ScrapeUnblocker key alone. There is no developer account, API key or partner approval to get on the target site.

New Retail and E-commerce Plugins

Eight retailers now have their own product search plugin. They all work the same way: you send a keyword, optionally a sort order and a price range, and you get a products array back. Pagination is built in, so you ask for the number of results you want instead of walking pages yourself.

What each one returns:

• Walmart - title, price and was-price, star rating, review count, brand, seller and availability.
• Target - TCIN, title, brand, price, image and a clean product URL.
• Best Buy - SKU, brand, model number, price and the regular price, rating and review count.
• The Home Depot - item ID, brand, model number, price, rating and review count.
• Etsy - title, price and currency, shop name, image and listing URL.
• SHEIN - goods ID, SKU, price, list price, discount, rating, review count, category and sold-out state.
• AliExpress - price, original price, rating and orders sold, with the currency set by the exit country.
• Zalando - SKU, brand, name, price and currency on the country storefront you choose.

Amazon got its own plugin at the end of August, with product and search endpoints across 20 regional marketplaces and prices in the right local currency. We covered it in How to Scrape Amazon Prices.

What You Can Build

• Price monitoring and repricing. Walmart’s was-price and Best Buy’s regular price tell you when something is marked down, not just what it costs today. Run a daily search for your product set and alert on new markdowns.
• MAP and brand compliance. Brands can check how resellers price their products across Walmart, Target, Best Buy and The Home Depot. Model numbers and SKUs make it much easier to match the same item across stores.
• Trend spotting. SHEIN’s discounts and review counts and AliExpress’s orders-sold figure show what is actually moving, which beats knowing what is merely listed. Etsy does the same for handmade and niche categories.
• Product datasets for AI. Brand, model, price and category from several retailers make a clean, structured set for training or testing product matching and classification models.

Here is a small markdown watcher for Walmart, using the parameters from the Walmart plugin docs:

import requests

r = requests.post(
    "https://api.scrapeunblocker.com/marketplace/walmart-search",
    params={"query": "laptop", "sort": "price_low", "max_results": 60},
    headers={"x-scrapeunblocker-key": "YOUR_API_KEY"},
    timeout=180,
)
r.raise_for_status()

for p in r.json()["products"]:
    price, was = p.get("price"), p.get("wasPrice")
    if price and was and price < was:
        drop = round(100 * (was - price) / was)
        print(f"{drop}% off  {price} (was {was})  {p['name'][:60]}  {p['url']}")

Fields with no value are left out of a product, so use .get() instead of indexing directly.

Marketplaces and Classifieds

Three second-hand and classifieds sites joined the catalogue:

• Vinted - resale listings with title, price and currency, brand, size, condition and a photo, on any regional marketplace.
• Craigslist - classifieds for any US metro and category, with title, price, location and image.
• leboncoin - France’s biggest classifieds site, with price, city, postcode, department and region, category, date, seller type (pro or private) and ad attributes.

What to build:

• Resale price guides. Pull Vinted listings for a brand, group them by size and condition, and publish a median price. Resellers and buyers both look for exactly this.
• Sourcing alerts. Watch Craigslist in a few metros for a category you resell, and get pinged when a new listing appears below your price threshold.
• Local market research. leboncoin’s seller type and location fields let you measure how much of a category is sold by professionals versus private sellers, region by region.

Travel: Southwest, Booking.com and Airbnb

Southwest Airlines fares fills a real gap, because Southwest does not sell its fares through the usual flight-search APIs. You pass IATA airport codes, dates and the number of adults, and choose fare_type=dollars or fare_type=points. The response lists every flight on the route with a price for each fare product (Wanna Get Away, Wanna Get Away Plus, Anytime and Business Select), including taxes and fees. The request, straight from the Southwest docs:

curl -X POST "https://api.scrapeunblocker.com/flights/southwest-quotes?origin=DAL&dest=HOU&depart_date=2026-10-20&return_date=2026-10-27&fare_type=dollars&adults=1" \
  -H "x-scrapeunblocker-key: YOUR_API_KEY"

Leave return_date off for a one-way search.

Booking.com hotels returns the properties for a location and dates: name, price for your stay, review score and count, full address, coordinates, a photo and badges. You set guests, rooms and currency.

Airbnb stays returns name, room type, price (USD by default) with any original struck-through price, rating, review count, badges such as Guest favourite, coordinates and a room URL. You can filter by dates, guests, price and room type.

What to build:

• Fare alerts and points value. Poll a Southwest route daily and alert when a fare drops. Pull the dollar and points fares for the same flight and you can see what a point is actually worth on that route.
• Hotel rate comparison. Hotels and revenue managers can track their own price and their competitors’ prices on Booking.com for the same dates.
• Short-term rental research. Airbnb coordinates, price and rating let you map nightly rates by neighbourhood before you buy or list a property.

Skyscanner car hire also got more precise. You can now pass Skyscanner’s own location IDs (pickup_entity and dropoff_entity) to search an exact place, and a place name that cannot be resolved returns an error instead of quietly searching somewhere else. Response headers tell you which location was actually searched. The original Skyscanner plugins post covers the basics.

Jobs: Indeed, Naukri and LinkedIn

• Indeed - title, company, location with a remote flag, salary text with a parsed minimum and maximum, job types, company rating and a snippet, on any Indeed country site.
• Naukri - India’s largest job board, with work mode, experience range, parsed salary, skills, company rating and posted date.
• LinkedIn jobs - title, company, location and posted date, filterable by LinkedIn’s guest filters (date posted, workplace, job type, experience and salary). A detail mode adds the full description, seniority, employment type, job function and industries.

What to build:

• Salary benchmarking. Parsed minimum and maximum salaries turn job listings into a pay dataset by role and city, with no regex on salary strings.
• Skills demand tracking. Naukri returns the skills for each posting. Count them week by week and you see which technologies employers ask for more often.
• Hiring signals for sales. A company posting five data engineering roles this month is probably about to buy data tools. Recruiters and B2B sales teams build lead lists from exactly this signal.

Real Estate and Cars

Property:

• Zillow - price, beds, baths, area, address, coordinates, home type and status, plus the Zestimate, rent Zestimate and tax-assessed value. For sale, for rent or sold.
• Redfin - price, beds, baths, square footage, lot size, MLS status, days on market and price per square foot. A detail mode adds listing agents, price history, the Redfin Estimate, tax records and nearby schools.
• Rightmove - UK sales and rentals with price, rent frequency, bedrooms, address, coordinates and the estate agent.
• ImmobilienScout24 - German listings with price and warm rent, rooms, living area, construction year, energy efficiency class and realtor.

Cars:

• mobile.de - make and model, price in EUR, first registration, mileage, power and fuel, with price, year, mileage and fuel filters.
• AutoScout24 - make, model, version, price, registration, mileage, fuel, gearbox, power and the seller’s name, type and location.

What to build:

• Comps and valuation. Redfin’s price history and days on market, or Zillow’s sold listings, give you comparable sales for a pricing model.
• Rental yield screens. Zillow returns the price and the rent Zestimate in one record, which is enough for a quick gross yield per listing. Rightmove sale and rental searches for the same area do the same job in the UK.
• Energy and renovation analysis. ImmobilienScout24’s energy class and construction year let you measure the price gap between efficient and inefficient homes in a German city.
• Used-car pricing. Price against mileage, year and power across mobile.de and AutoScout24 is the core of a used-car valuation model. Dealers also use the same data to watch competitor stock.

Social, Video and Ads

TikTok now has a full plugin with five endpoints:

• tiktok-profile - exact follower, like and video counts, bio and flags, plus up to 200 of the creator’s newest videos.
• tiktok-video - one video or photo post with exact plays, likes, comments, shares, saves and reposts, hashtags, music, media URLs, subtitle tracks and an optional transcript.
• tiktok-hashtag - total views and video count for a tag, with its top videos.
• tiktok-search - the videos TikTok ranks for a keyword, with a selectable region.
• tiktok-comments - a post’s comments with author, likes and replies.

“Exact” is the important word. TikTok shows 1.6M on screen, and the plugin returns the real number underneath. Two practical notes from the TikTok docs: search and comments are slower, at roughly 20-45 seconds a call, and media URLs expire after a few hours, so download what you need in the same run.

YouTube gained search and comments plugins. Search returns the video ID, title, channel, view count as a number, duration in seconds, publish age and badges, with type and sort filters. Comments returns the author, text, like and reply counts and pinned or hearted flags, sorted by top or newest. The channel endpoint also started returning exact view, like and comment counts, which we covered in How to Scrape YouTube Channel Data as JSON.

Threads profiles come back with bio, follower count, thread count and recent posts with likes, comments, reposts and shares.

Meta Ad Library returns an advertiser’s Facebook and Instagram ads: ad copy, headline, call to action, format, image and video creative URLs, the landing page, platforms and run dates. Through Meta’s official Ad Library API, full data is limited to political and social-issue ads, so this is the practical way to see what an ordinary commercial advertiser is running. This snippet prints how long each active ad has been live:

from datetime import date
import requests

r = requests.post(
    "https://api.scrapeunblocker.com/ads/meta-ad-library",
    params={"advertiser": "Nike", "country": "US", "active_status": "active", "max_ads": 30},
    headers={"x-scrapeunblocker-key": "YOUR_API_KEY"},
    timeout=180,
)
r.raise_for_status()

for ad in r.json()["results"]:
    if not ad.get("started_running"):
        continue
    days = (date.today() - date.fromisoformat(ad["started_running"])).days
    print(f"{days:>4} days  {ad['display_format']:<9} {ad.get('cta_text')}  {ad.get('link_url')}")

What to build:

• Influencer vetting. Compare a creator’s follower count with the exact plays and likes on their recent videos. A big audience with weak engagement shows up at once.
• Brand and sentiment monitoring. Pull the comments on TikTok and YouTube videos that mention your product and run them through a sentiment model.
• Content research. TikTok and YouTube search show what ranks for a topic right now, with the engagement numbers to back it up.
• Competitor ad research. An ad that has run for 60 days is usually one that works. Sort a competitor’s ads by run time, study the long-runners, and note where they send the click.

Local Businesses, Apps and Reviews

• Google Maps Place Details - the full record for one place: name, address, phone, website, rating, categories, coordinates, plus code and today’s hours with an open-now flag. It pairs well with the Google Local plugin, which finds the businesses for a search in the first place.
• Yelp - search US businesses by term and city, or fetch one by URL. You get rating, review count, price tier and categories, and detail mode adds phone, website, hours and popular dishes.
• Trustpilot - a company’s TrustScore, its 1-5 star breakdown, and each review with its text, verified flag, author country and the company’s reply.
• Google Play - one app’s title, developer, score, full rating histogram, installs, price, content rating, version and what’s-new notes, or a ranked app list for a search.
• Product Hunt - the daily leaderboard, keyword search, topics or a single product, with votes, comments, rank, topics and makers.

What to build:

• Local lead generation. Search Google Local for a category and city, then add Yelp’s price tier and review counts to rank the prospects. You end up with a qualified list with phones and websites.
• Reputation dashboards. Track a company’s Trustpilot star distribution and how often it replies to reviews, next to its competitors.
• App store intelligence. Watch competitors’ rating histograms and update dates on Google Play. A drop in five-star share right after a release is a clear signal.
• Launch tracking. Product Hunt topics and leaderboards show which products launch in your category and how the community reacts.

Google Images Got Better

The Google Images plugin launched in mid-September and has improved twice since:

• A new pages parameter fetches up to 5 result pages of about 100 images each. Each page is billed as one request, and pagesFetched in the response tells you how many were used. max_results now goes up to 500.
• The Google market now follows proxy_country automatically, so results match the country you picked. gl stays available as an optional override.
• Every result now carries googleImagesUrl, the Google Images viewer link for that image.

Together these make the plugin much more useful for building image datasets for AI, checking where your product photos appear, and any research that needs more than one page of results.

Platform Improvements

Browser steps and element discovery. getPageSource can now click, type, select, scroll and wait in a real browser before it returns the HTML, and list_elements=true gives you a page’s interactive elements with ready-to-use selectors. Interact With a Page, Then Scrape It walks through it.

ChatGPT connector. The hosted MCP server now works as a custom connector in ChatGPT on the web, with a step-by-step setup guide. ChatGPT gets four tools: fetch a page’s HTML, fetch parsed JSON, run a Google search and list a page’s interactive elements. You need a paid ChatGPT plan with developer mode turned on.

Clearer errors. Two changes make failures easier to handle:

• An unreachable target site now returns 523 instead of 403. A 523 is not a bot block and is never billed. Retry it after a pause.
• A page that does not exist on the target (404 or 410) now comes back as 200 with an X-Origin-Status header carrying the target’s real status. Retrying will not change it, so drop or fix the URL.

The errors reference now lists every status with whether it is billed, and the handling failures guide has a recovery playbook for each. If your retry logic only looked at 403, update it along these lines:

import time
import requests

RETRYABLE = {500, 502, 503, 504, 523}

def fetch(url, key, max_attempts=4):
    for attempt in range(max_attempts):
        r = requests.post(
            "https://api.scrapeunblocker.com/getPageSource",
            params={"url": url},
            headers={"x-scrapeunblocker-key": key},
            timeout=125,
        )
        if r.status_code == 200:
            if r.headers.get("X-Origin-Status"):
                return None  # the target says this URL is dead: drop or fix it
            return r.text
        if r.status_code in RETRYABLE:
            time.sleep(5 * 2 ** attempt)
            continue
        r.raise_for_status()
    return None

Plans. Every plan now has two limits per API key, applied together: 10 to 50 concurrent requests and 2 to 10 requests per second on Pay as you go, Starter and Pro, and 100 to 400 concurrent requests and 20 to 50 requests per second on Business, Scale and Ultimate. A request over either limit gets an immediate 429, which is not billed, so back off and retry it. Scale and Ultimate include Slack support. Details are on the pricing page.

Docs. The documentation moved to docs.scrapeunblocker.com. Old links redirect to the same page, so bookmarks keep working.

Where to Start

• Browse the scraper catalogue. Each page shows the fields, a sample request and the public benchmark for that plugin.
• Try a plugin without code in the dashboard. Fill in a few fields and look at the JSON before you write anything.
• Follow the changelog for the smaller updates between roundups.

FAQ

Do I need an account or API key on the target sites? No. Every plugin runs with your ScrapeUnblocker key alone. There is no Walmart, Meta, TikTok or Skyscanner developer account involved.

What is the difference between a plugin and the regular scraping endpoint? getPageSource returns the HTML (or parsed JSON) for any URL, and you extract the data yourself. A plugin takes plain inputs such as a keyword or a route and returns the fields already parsed, with pagination handled for you.

Where do I find the exact parameters for a plugin? On that plugin’s page in the scraper catalogue and in the plugin docs. Parameters go in the query string of a POST request.

Wrapping Up

That is the month. If one of these plugins fits a project you have been putting off, the fastest test is a single request from the ScrapeUnblocker dashboard. We will be back with the next roundup once there is enough new to report.