DEV Community: Azzam

How to Scrape Google Maps (With Emails) in Python (2026 Guide)

Azzam — Thu, 15 Oct 2026 14:20:03 +0000

Google Maps is the biggest directory of local businesses there is, and if you sell to dentists, plumbers or restaurants it's where your lead list lives. The problem is that copying listings by hand is slow, and Maps never shows an email address. This guide shows how to scrape Google Maps from Python, get one clean row per business, and add the emails each business publishes on its own website.

Quick start

pip install apify-client

from apify_client import ApifyClient

client = ApifyClient("YOUR_APIFY_TOKEN")

run = client.actor("automly/google-maps-scraper").call(run_input={
    "searchStringsArray": ["dentists in Chicago"],
    "maxCrawledPlacesPerSearch": 50,
    "includeContacts": True,
})

for place in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(place["title"], "|", place.get("phone"), "|", place.get("emails"))

That's a working Google Maps scraper in Python. Your token is in the Apify Console under Settings > API & Integrations. Not using Python? The same actor works as a plain Google Maps scraper API: a POST to Apify's run-sync-get-dataset-items endpoint with the same JSON input returns the rows as JSON.

Why not the official API?

Google does have a Google Maps API for this: the Places API (New). It works, but two things make it a poor fit for lead lists. It has no email field at all, and one Text Search query stops at 60 results across all pages. The phone, website, rating and opening hours fields also sit in the Enterprise tier, so asking for them bills the more expensive SKU.

Since March 1, 2025, Google no longer gives a $200 monthly credit. Each SKU now has its own free monthly cap instead. Prices below are the first volume tier from Google's pricing page, checked October 2026:

Option	Access needed	Cost	What you get
Places API (New), Text Search Pro	Google Cloud project, billing account, API key	$32 per 1,000 requests, 5,000 free per month	Name, address, category, location. No phone, website or rating
Places API (New), Text Search Enterprise	Same	$35 per 1,000 requests, 1,000 free per month	Adds phone, website, rating, review count, hours. Up to 20 places per request, 60 per query. No email
Places API (New), Place Details Enterprise	Same	$20 per 1,000 requests, 1,000 free per month	The same Enterprise fields for one place per request. No email
Google Maps Scraper (this guide)	Free Apify account	$3.00 per 1,000 places, plus $0.01 per run and platform usage	Name, category, phone, website, address, rating, reviews, hours and more. Optional emails and social profiles from the business's website

If you need a handful of lookups inside an app, the official API is fine. If you need every dentist in a city with an email column, you need something else.

Step 1: Pick what and where to search

There are two ways to say where. You can put the city in the query, like plumbers in Denver, or keep the query short and set locationQuery:

run_input = {
    "searchStringsArray": ["plumber", "electrician"],
    "locationQuery": "Austin, TX",
    "maxCrawledPlacesPerSearch": 100,
    "language": "en",
}

locationQuery takes a city, a state, a postal code or a country, and only places inside that area are kept. Each query in searchStringsArray runs separately, and maxCrawledPlacesPerSearch caps each one (the default is 100).

You can also skip searching and pass Google Maps links in startUrls or place IDs in placeIds if you already know which businesses you want.

Step 2: Turn on emails and social profiles

This is the part people mean when they search for a "Google Maps email scraper". Set includeContacts to True and the scraper visits each place's own website and adds three fields:

emails: addresses found on the site
phones: phone numbers found on the site (often more than the one on Maps)
socialProfiles: Facebook, Instagram, LinkedIn, X, YouTube, TikTok and Pinterest links

To be clear about where these come from: Google Maps itself doesn't publish emails. They come from the business's website. So places without a website get no email, and some websites simply don't list one. In a recent run of 15 Austin dentists, 14 came back with an email or a social profile.

Step 3: Filter before you pay for rows

Filters run before a place is saved, so skipped places aren't charged:

run_input.update({
    "placeMinimumStars": "four",       # also: "threeAndHalf", "fourAndHalf", ...
    "placeMinimumReviews": 20,
    "website": "withWebsite",          # or "withoutWebsite", "allPlaces"
    "skipClosedPlaces": True,
})

A tip if you sell web design: set website to "withoutWebsite". Those businesses are your leads, and you call them instead of emailing.

Step 4: Read the output

Here's a real row, trimmed and anonymized:

{
  "title": "Example Hills Dentistry",
  "categoryName": "Dentist",
  "address": "Example Hills Dentistry, 1200 W Example Blvd #207, Austin, TX 78704",
  "city": "Austin",
  "postalCode": "78704",
  "website": "https://www.example-dentist.com/",
  "phone": "(512) 555-0147",
  "phoneUnformatted": "+15125550147",
  "totalScore": 5,
  "reviewsCount": 354,
  "openingHours": [{"day": "Friday", "hours": "8 AM-4 PM"}],
  "placeId": "ChIJ...",
  "url": "https://www.google.com/maps/place/?q=place_id:ChIJ...",
  "emails": ["hello@example-dentist.com"],
  "phones": ["+15125550147", "+17375550192"],
  "socialProfiles": {
    "facebook": ["https://www.facebook.com/example"],
    "instagram": ["https://www.instagram.com/example"]
  }
}

placeId stays the same between runs, which makes it a good key when you refresh a list and want to spot new businesses.

Step 5: Save to CSV

import csv

rows = client.dataset(run["defaultDatasetId"]).iterate_items()
fields = ["title", "categoryName", "phone", "website", "address",
          "totalScore", "reviewsCount", "emails", "placeId"]

with open("leads.csv", "w", newline="", encoding="utf-8") as f:
    writer = csv.DictWriter(f, fieldnames=fields, extrasaction="ignore")
    writer.writeheader()
    for place in rows:
        place["emails"] = "; ".join(place.get("emails") or [])
        writer.writerow(place)

If you prefer pandas, pd.DataFrame(client.dataset(...).list_items().items) gets you there too.

What else can you scrape?

The same actor can add more to each place:

maxReviews: the newest (or most relevant) reviews per place, with text, rating, date and the owner's reply. reviewsStartDate limits them to recent ones.
maxImages: photo links with the date each was added.
scrapePlaceDetailPage: popular times, review counts per star, review topics and "people also search for".
categoryFilterWords: keep only places whose Google category matches, say, italian restaurant.

Scaling

One Google Maps search shows around 200 places, and a big city has far more. Set enableGridSearch to True and the area is split into smaller cells that are searched one by one and merged, so a single query can return thousands of places. gridCellKm sets the cell size: 5 for a quick pass, 2 for a thorough one, 1 for the most complete list. For a city that isn't recognized automatically, pass customBounds with south, west, north and east.

For a list you keep fresh, save the input as a task in Apify and put it on a weekly schedule.

No code? Use the Apify Console

Open the actor page, click Try for free, type your searches and location, flip Add emails and social profiles, and click Start. The results show in tables like Overview, Contacts and Website contacts, and download as CSV, Excel or JSON. Website contacts puts the name, website, emails, phones and social profiles side by side, and it's the one I'd hand to a sales team. Google Sheets, Zapier, Make and n8n can pick up each run.

What it costs

Right now places cost $3.00 per 1,000, plus $0.01 per run start and the run's Apify platform usage. Contact details have no separate charge today.

That changes on October 17, 2026, when contact details get their own price, charged only for places where the website gave at least one email, phone or social profile:

Apify plan	Per 1,000 places with contacts found
Free	$1.80
Starter	$1.60
Scale	$1.20
Business and up	$0.85

Every Apify account gets $5 of free usage per month, which covers roughly 1,500 plain places. The Pricing tab on the actor page always has the current numbers.

Responsible use

Stick to public business data. Some of it, like an owner's name or a personal email, can still be personal data under GDPR or CCPA, and cold outreach has its own rules such as CAN-SPAM. Follow Google's terms and ask a lawyer if you're unsure.

Links

If you want a lead list by this afternoon, try Google Maps Scraper on Apify. The free $5 a month is enough to test it on your own city.

How to Scrape LinkedIn Profiles Without Logging In (Python, 2026 Guide)

Azzam — Wed, 14 Oct 2026 14:20:04 +0000

How to Scrape LinkedIn Profiles Without Logging In (Python, 2026 Guide)

Building a lead list or a candidate shortlist on LinkedIn by hand works for 20 people and falls apart at 500. This guide shows how to scrape LinkedIn profiles from Python by job title, company, location, school or name, without a LinkedIn account or Sales Navigator. You get one row per person with the name, headline, company, location and profile link.

Here's the whole LinkedIn scraper in Python. Install the client with pip install apify-client, paste your Apify token, and run:

from apify_client import ApifyClient

client = ApifyClient("YOUR_APIFY_TOKEN")

run = client.actor("automly/linkedin-people-search-scraper").call(run_input={
    "jobTitles": ["Product Manager"],
    "companies": ["Stripe"],
    "locations": ["San Francisco"],
    "maxResults": 50,
})

for person in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(person["name"], "|", person["headline"], "|", person["linkedinUrl"])

Below I go through the inputs, the output, the limits and the price.

Why not the official API?

LinkedIn does have APIs, but none that an ordinary developer can sign up for lets you search for people. I checked the official docs in October 2026:

Option	Access needed	Cost	What you get
Sign In with LinkedIn (OpenID Connect)	Self-serve app, and each person must sign in to it	No fee at present (API Terms, 8.2)	The signed-in member's own name, photo and email
Profile API	Developers approved by LinkedIn	No fee at present (same API Terms)	Another member's profile only if you already have their Person ID, which comes from other limited-access APIs, and you may never store it
Sales Navigator Application Platform (SNAP)	Approved SNAP partner	Not published	Sales Navigator data, such as matched public profile info, inside a partner's app
This scraper	An Apify account	$1.50 per 1,000 profiles	People search by title, company, place, school or name

According to LinkedIn's Getting Access page, the only permissions open to every developer are sign-in (profile, email) and posting on the member's behalf. Everything else goes through a partner program that LinkedIn has to approve. And nothing in that permissions list lets you find other members by job title or company.

Step 1: Get your Apify token

Create a free Apify account, then copy your API token from your Apify Console settings. Every account gets $5 of free usage each month, which is plenty for testing.

Step 2: Choose your filters

The actor reads these input fields:

Field	What it does
`jobTitles`	List of titles, e.g. `["Product Manager", "Head of Sales"]`
`companies`	Company names or LinkedIn company links
`locations`	Cities, regions or countries
`schools`	Universities or schools
`searchQuery`	Any other words the profile should mention, e.g. `"fintech"`
`firstName`, `lastName`	Find a person by name
`maxResults`	Stop after this many profiles in the whole run (default 100)
`maxResultsPerSearch`	Cap each search, e.g. 50 per city
`fullProfiles`	Also read each profile page for more fields (default off)

Every combination of one title, one company, one location and one school counts as one search. Two titles and three cities give you six searches, and a person found by more than one of them is saved once.

To find one person by name:

run_input = {"firstName": "Jane", "lastName": "Doe", "companies": ["Stripe"]}

The locations filter is strict. People whose profile shows another place, or no place at all, are left out. Add the country ("Berlin, Germany") to also keep people who list only their country.

Step 3: Read the results

Each row from the search looks like this (name and link anonymized):

{
  "linkedinUrl": "https://www.linkedin.com/in/jane-doe-123",
  "publicIdentifier": "jane-doe-123",
  "name": "Jane D.",
  "headline": "Product Manager @ Stripe",
  "company": "Stripe",
  "school": "University of Michigan",
  "location": "San Francisco Bay Area",
  "connections": "500+",
  "query": "\"Product Manager\" \"Stripe\" \"San Francisco\""
}

query tells you which search found the person, which helps when one run covers many cities or titles.

Step 4: Add profile details

Set "fullProfiles": True and each row gets extra fields: about, countryCode, photo, followers, badges, currentCompany, currentPositions, pastCompanies, education, recentPostUrls and profileVisibility. A typical row looks like this (trimmed, anonymized):

{
  "name": "John Smith",
  "headline": "Data Scientist at Shopify",
  "location": "Toronto, Ontario, Canada",
  "countryCode": "CA",
  "followers": 410,
  "connections": "500+",
  "currentCompany": "Shopify",
  "education": [{"name": "University of Waterloo", "startDate": "2012", "endDate": "2016"}],
  "experience": [],
  "profileVisibility": "limited"
}

See the empty experience? That's normal for most people, which brings me to the limits.

Step 5: Know what a logged-out profile hides

LinkedIn decides how much a visitor who isn't logged in can see, and the scraper can only return that. Plan around it:

Most profiles are limited. profileVisibility is full, limited or hidden. Work history (experience) and jobTitles only come back for full profiles, which are mostly creators, founders and executives. In the two detail runs I checked, 22 of 35 people were limited, 13 were hidden and none were full. On limited profiles, about was filled in on only 9 of 22.
Hidden profiles keep only the search data. You still get the name, headline and, when listed, company, location, school and connections.
Some last names are cut to an initial. About 1 in 10 rows in my runs had a name like "Jane D." LinkedIn shows those members that way to logged-out visitors, and that's all the scraper receives.
There are no emails or phone numbers.
Filters match words on the profile, not LinkedIn's internal fields, so someone who mentions Stripe as a past client can show up. In the runs I checked, 170 of 173 rows had the searched company in the headline or company field. Filter the company column if you need exact matches.
Seniority, function and company-size filters need an account. Put words like "Senior" or "Director" into the job title instead.

Step 6: Save to CSV

import csv

items = list(client.dataset(run["defaultDatasetId"]).iterate_items())
fields = ["name", "headline", "company", "location", "school",
          "connections", "linkedinUrl", "query"]

with open("linkedin_profiles.csv", "w", newline="", encoding="utf-8") as f:
    writer = csv.DictWriter(f, fieldnames=fields, extrasaction="ignore")
    writer.writeheader()
    writer.writerows(items)

extrasaction="ignore" drops any fields you didn't list. With pandas, pd.DataFrame(items).to_csv("linkedin_profiles.csv", index=False) works too.

What else can you scrape?

The same code works for my other LinkedIn scrapers. Change the actor name and the input:

LinkedIn Profile Scraper + Email Finder: profiles from links or names, plus work emails. Feed it the linkedinUrl column from this guide.
LinkedIn Company Employees Scraper: the people who work at a company.
LinkedIn Posts Scraper: posts from people and company pages.
LinkedIn Post Search Scraper: posts about any topic.

Scaling

One search returns up to about 1,000 profiles. To get more, split it: add more cities, titles or schools, since each combination is its own search. Use maxResultsPerSearch when you want an even spread, for example 100 per city. The run report shows how many profiles each search found, and a search that stops near 1,000 has more to give if you split it further.

In my tests a run with several searches returned about 1,500 profiles per minute, and a single search about 500.

No code? Use the Apify Console

Open the actor page, click Try for free, type the titles, companies and cities one per line, and click Start. Results download as Excel, CSV, JSON, HTML or XML, or go straight to Google Sheets. Save the input as a task and put it on a schedule if you want the same search every week.

Cost

You pay per profile saved: $1.50 per 1,000 profiles from the search, and $2.50 per 1,000 with profile details. Limited and hidden profiles are charged at the lower price even when details are on. Each run also has a start fee of a fraction of a cent. The price is the same on every Apify plan.

The $5 of free monthly usage covers about 3,300 search profiles, or at least 2,000 with details. Check the Pricing tab on the actor page before a big run, since prices can change.

Responsible use

This is public data that LinkedIn shows to logged-out visitors, but a profile is still personal data. Under the GDPR and CCPA you need a lawful reason to collect it, should keep only what you need, and have to honor deletion requests. LinkedIn's User Agreement also restricts scraping, so read it and talk to a lawyer before you build a product or an outreach campaign on this data.

Links

If you want a LinkedIn profile scraper you can call from Python without logging in, try LinkedIn People Search Scraper on Apify. Your first runs fit in the free $5.

How to Scrape Instagram Comments in Python Without Logging In (2026 Guide)

Azzam — Tue, 13 Oct 2026 14:20:03 +0000

You have a few Instagram posts and you want every comment under them in a spreadsheet: username, text, likes and date. Maybe it's a giveaway, maybe a client wants to know what people said about a launch. This guide shows how to scrape Instagram comments in Python without an Instagram account, using a hosted Instagram comment scraper on Apify.

Quick start

Install the client:

pip install apify-client

Then run this with your Apify API token:

import csv
from apify_client import ApifyClient

client = ApifyClient("YOUR_APIFY_TOKEN")

run = client.actor("automly/instagram-comments-scraper").call(run_input={
    "postUrls": [
        "https://www.instagram.com/p/DdG4RIxIPyf/",
        "https://www.instagram.com/reel/DdHyaYAifb6/",
    ],
    "resultsPerPost": 500,
})

comments = list(client.dataset(run["defaultDatasetId"]).iterate_items())

with open("instagram_comments.csv", "w", newline="", encoding="utf-8") as f:
    writer = csv.DictWriter(
        f,
        fieldnames=["postUrl", "ownerUsername", "text", "likes", "createdAt"],
        extrasaction="ignore",
    )
    writer.writeheader()
    writer.writerows(comments)

print(f"Saved {len(comments)} comments")

That's the whole thing. You get up to 500 comments from each post, newest first, in instagram_comments.csv. No Instagram login, cookies or Meta developer app.

Why not the official API?

Meta does have an Instagram API, and if you only need comments on posts from accounts you manage, it's a fine choice. For anyone else's posts it doesn't help much.

Option	Access needed	Cost	What you get
Instagram API, `/comments` edge	An Instagram professional (Business or Creator) account, a token from the user who created the post, `instagram_business_basic` + `instagram_business_manage_comments` permissions. App Review and Business Verification for accounts you don't own or manage	No fee listed by Meta	Comments on your own media. Up to 50 per query, top-level only by default, no date filter
Business Discovery	Your own professional account and a Meta app	No fee listed by Meta	Follower count, media count and `comments_count` for other professional accounts. Not the comment text, and age-gated accounts return nothing
Instagram Basic Display API	n/a	n/a	Shut down on 4 December 2024
Instagram Comments Scraper	An Apify account	$0.40 per 1,000 comments	Comments on any public post or reel, with likes, date and commenter

So the official route gives you comment counts for other accounts, and comment text only for your own. If the posts belong to someone else, you need a different tool.

Step 1: Get an Apify token

Create a free account at apify.com (no credit card). Your token is in Apify Console under Settings > API & Integrations. Put it in an environment variable rather than in your code:

import os
from apify_client import ApifyClient

client = ApifyClient(os.environ["APIFY_TOKEN"])

Step 2: Collect the post links

postUrls takes post links (https://www.instagram.com/p/...) and reel links (https://www.instagram.com/reel/...). Links copied from the app with extra parameters on the end work too. You can pass as many as you want in one run, and every comment row tells you which post it belongs to.

Step 3: Choose how many comments

resultsPerPost is the most comments to get from each post. The default is 100 and it goes up to 100,000. Comments come newest first, and the scraper stops when it has that many or the post runs out.

One thing to know before you compare numbers: the count Instagram shows on a post includes replies to comments. This scraper collects the comments themselves, not the replies under them, so you'll usually get fewer rows than the number on the post.

Step 4: Only get new comments

This one matters when you follow a post over several days: without it, every run pays again for comments you already have. commentsPostedAfter keeps only comments posted on or after a date (UTC if you don't give an offset):

run_input = {
    "postUrls": ["https://www.instagram.com/p/DdG4RIxIPyf/"],
    "resultsPerPost": 10000,
    "commentsPostedAfter": "2026-10-01",
}

It also takes a time, like 2026-10-01T12:00:00Z. Because comments arrive newest first, the run stops as soon as it reaches older ones.

Step 5: Read the output

Each row is one comment. Here's a trimmed one (names changed):

{
  "id": "17995635840044270",
  "text": "Amazing shot!",
  "createdAt": "2026-10-02T10:24:54Z",
  "likes": 3,
  "ownerUsername": "travel.with.jane",
  "ownerId": "1234567890",
  "ownerVerified": false,
  "ownerProfilePicUrl": "https://scontent.cdninstagram.com/v/...jpg",
  "postShortCode": "DdG4RIxIPyf",
  "postId": "3983118388705688735",
  "postUrl": "https://www.instagram.com/p/DdG4RIxIPyf/",
  "source": "https://www.instagram.com/p/DdG4RIxIPyf/",
  "fetchedAt": "2026-10-02T10:31:42Z"
}

source is the exact link you passed in, and postUrl and postShortCode identify the post. If you'd rather work in pandas, the same list goes straight into a DataFrame:

import pandas as pd

df = pd.DataFrame(comments)
df["createdAt"] = pd.to_datetime(df["createdAt"])
df.to_excel("instagram_comments.xlsx", index=False)
print(df.groupby("postShortCode")["likes"].sum())

What else can you scrape?

Most comment jobs start before the comments: you need to find the posts, or you want to know more about who wrote them.

Instagram Posts & Reels Scraper gets a profile's posts and reels, so you can pick the ones worth scraping comments from.
Instagram Post Scraper returns full details of posts and reels from their links.
Instagram Profile Scraper gives bio, followers, following and post count for usernames, such as your commenters.
Instagram Scraper does posts, profiles, hashtags and comments in one tool, if you'd rather learn how to scrape Instagram data from a single place.

Scaling up

Put all your links in one run instead of starting a run per post. In my test, 30 posts gave 6,000 comments in 41 seconds.

For monitoring, save the input as a task in Apify Console, add a daily schedule, and move commentsPostedAfter forward each time (or set it from your own code before calling). Results can go to Google Sheets, Slack, Make or Zapier through Apify integrations.

Each run also saves a short report: how many posts and comments it covered, and any link it couldn't scrape with the reason, such as a deleted or private post. Read it from Python:

store = client.key_value_store(run["defaultKeyValueStoreId"])
report = store.get_record("report-summary")
print(report["value"])

You can also set a maximum cost on a run in Console, and the run stops when it reaches it.

No code? Use the Console

Open Instagram Comments Scraper and click Try for free.
Paste post or reel links into Post or reel URLs, one per line.
Set Comments per post and, if you want, Comments posted after.
Click Start, then export the Instagram comments as CSV, Excel, JSON, HTML or XML.

What it costs

You pay per comment you get: $0.40 per 1,000 comments, the same on every Apify plan, plus $0.00005 when a run starts. 10,000 comments cost about $4. Every Apify account gets $5 of free usage each month, which covers roughly 12,000 comments at that price.

Responsible use

Only scrape public posts, and only for a real reason. Comments can contain personal data that laws like GDPR and CCPA protect, so follow Instagram's terms and ask a lawyer if you're not sure. The tool isn't affiliated with Instagram or Meta.

How to Scrape the Facebook Ad Library in Python (2026 Guide)

Azzam — Mon, 12 Oct 2026 13:49:05 +0000

The Meta Ad Library shows every active ad running on Facebook, Instagram and the rest of Meta's apps, but the website only lets you scroll through it one card at a time. This guide shows how to scrape Facebook ads from the library with Python, using a Facebook Ad Library scraper on Apify, and get one clean row per ad: text, headline, images, videos, landing page, call to action, dates, platforms and the advertiser's Page. You don't need a Facebook account or a Meta developer app.

Quick start

pip install apify-client

from apify_client import ApifyClient

client = ApifyClient("YOUR_APIFY_TOKEN")

run = client.actor("automly/facebook-ad-library-scraper").call(run_input={
    "searchTerms": ["meal kit"],
    "country": "US",
    "activeStatus": "active",
    "maxAds": 100,
    "outputFormat": "clean",
})

for ad in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(ad["pageName"], ad["startDate"], ad["activeDays"], ad["linkUrl"])

Get your token from the Apify Console under Settings > API & Integrations. That's the whole thing. The rest of this post covers the inputs, the output, and when Meta's own API is the better choice.

Why not the official Meta Ad Library API?

Meta does have an Ad Library API. Before you can call it you have to confirm your identity and location at Facebook.com/ID (Meta says this "can take a few days"), create a Meta for Developers account and set up an app.

The bigger issue is what it returns. Per Meta's own page, the API covers ads about social issues, elections or politics from anywhere in the world over the past 7 years, plus ads of any type delivered to the EU or UK during the past year. The EU part comes from the Digital Services Act: in 2023 Meta expanded the library to archive every ad shown to people in the EU for a year. The ads_archive reference spells out the catch: ads that did not reach any EU location only come back if they are political or issue ads. So a US-only ad for a skincare brand isn't in the API at all.

The fields are mostly text lists (bodies, link titles, captions, descriptions), dates, the Page, platforms and an ad_snapshot_url that opens the ad. Spend, impressions and demographics are only filled for political and issue ads; reach and targeting only for EU/UK ads. There is no field for image or video files, the CTA button or the landing page URL.

Option	Access needed	Cost	What you get
Ad Library website	None. Anyone can view and search it	Free	All active ads across Meta apps, one at a time, no export
Meta Ad Library API	ID confirmation, developer account, app	No fee listed by Meta	Political/issue ads worldwide (7 years); other ads only if delivered to the EU/UK (past year); text fields and a snapshot link
Facebook Ads Scraper	Free Apify account	$0.70 per 1,000 ads on the Free plan, less on paid plans	Any ad the library shows, in any country, with media links, landing page, CTA and Page data

If you only research political ads, or only care about EU ads, the official API is a good fit. For commercial ads elsewhere, a Facebook ads library scraper is the practical route.

Step 1: Pick your input

There are three ways in, and you can mix them in one run.

Keywords. searchTerms takes a list, and each term runs as its own search. Set searchType to keyword_exact_phrase if you want the words in order.

Advertiser Pages. Put Page links in urls to get every ad one advertiser runs:

run_input = {
    "urls": [{"url": "https://www.facebook.com/nike"}],
    "country": "ALL",
    "activeStatus": "all",
    "maxAds": 1000,
    "outputFormat": "clean",
}

Ad Library search links. If you already set up a search on the Ad Library website, copy the link from the address bar and paste it into urls. Its own filters are used as they are. A single ad link (https://www.facebook.com/ads/library/?id=...) works too.

Step 2: Set the filters

country: a two-letter code (US, GB, DE), a country name, or ALL. Defaults to US.
activeStatus: active (default), inactive or all.
mediaType: all, image, video, meme or none (text only).
adType: all, or a special category such as political_and_issue_ads or housing_ads.
startDateMin / startDateMax: only ads that started in that window, as YYYY-MM-DD.
maxAds: cap per search term or URL. Default 100, maximum 10,000.

Always pass "outputFormat": "clean" from code. Runs that leave it out get the older field layout so existing integrations don't break.

Step 3: Read the output

Here is a real row from a netflix keyword search, trimmed (media links shortened, emoji removed from the ad text):

{
  "adArchiveId": "742409480628500",
  "adLibraryUrl": "https://www.facebook.com/ads/library/?id=742409480628500",
  "isActive": true,
  "startDate": "2023-03-21",
  "endDate": null,
  "activeDays": 1145,
  "adFormat": "video",
  "mediaType": "video",
  "bodyText": "Wow! These are the most comfortable headphones for relaxing...",
  "headline": "Get Your CozyBand Today",
  "linkUrl": "https://www.cozyband.com/products/cozyband",
  "linkDomain": "cozyband.com",
  "ctaText": "Shop now",
  "videos": [{"hdUrl": "https://video.xx.fbcdn.net/...mp4", "sdUrl": "...", "previewImageUrl": "..."}],
  "platforms": ["facebook", "instagram", "audience_network", "messenger"],
  "pageName": "CozyBand",
  "pageUrl": "https://www.facebook.com/getcozyband/",
  "pageCategory": "Electronics",
  "searchTerm": "netflix"
}

Image ads carry imageUrls, and carousels come with a cards list where each card has its own headline, link, CTA and image. activeDays is the column I sort by first. An ad that has kept running for months is usually one that pays for itself.

Step 4: Save it to CSV

import csv

cols = ["pageName", "headline", "bodyText", "linkUrl", "ctaText",
        "startDate", "activeDays", "mediaType", "adLibraryUrl"]

with open("ads.csv", "w", newline="", encoding="utf-8") as f:
    writer = csv.DictWriter(f, fieldnames=cols, extrasaction="ignore")
    writer.writeheader()
    for ad in client.dataset(run["defaultDatasetId"]).iterate_items():
        writer.writerow(ad)

You can also skip the code and download the dataset from the Console as CSV, Excel, JSON, HTML or XML.

What else can you scrape?

There are three optional switches, all off by default. Set includeAdDetails to true and ads shown in the EU or UK come back with their reach, targeted ages, gender and places, and who paid for them. Political and issue ads get their audience breakdown and the spend range Meta publishes.

includePageTransparency adds when the advertiser's Page was created, its past names and how many ads it runs. includeAdvertiserDetails adds the advertiser's Instagram account and follower count, verified badge and About text.

A note on spend, since people ask: Meta only publishes spend ranges for political and issue ads. For ordinary commercial ads no tool reading the Ad Library can give you a real spend figure.

Scaling up

Add more terms or Page links to the same run, and raise maxAds up to 10,000 per search. For ongoing competitor tracking, save your input as a task in the Console and put it on a schedule. Each run can push its dataset to Google Sheets, Slack or a webhook, and n8n, Make and Zapier can start runs too.

No-code option: the Apify Console

Open the actor page and click Try for free.
Fill in Search Terms or paste links into Ad Library URLs.
Pick Country, Ad Status and Max Ads per Search. Output format is already set to Clean.
Click Start, then export the results.

What it costs

You pay per ad saved: $0.70 per 1,000 ads on the Free plan, with lower rates on paid Apify plans. Each run also has a $0.05 start fee at the default settings. Every Apify account gets $5 of free usage each month, which covers roughly 7,000 ads.

Responsible use

Stick to public ad data and a clear purpose for collecting it. Advertiser Pages and political ad disclaimers can contain personal data, which laws like GDPR and CCPA may protect, so follow Meta's terms and check with a lawyer if you're unsure. The scraper isn't affiliated with Meta.

Links

If you want your competitors' Facebook and Instagram ads in a spreadsheet today, try the Facebook Ads Scraper on Apify. The free monthly credit covers a few thousand ads, which is plenty for a first look at three or four brands.

How to Scrape Amazon Product Data in Python (2026 Guide)

Azzam — Mon, 12 Oct 2026 13:48:45 +0000

If you sell on Amazon, track competitors or research a niche, you probably want the same few numbers for a list of products: price, rating, Best Sellers Rank and some idea of how many units move each month. This guide gets you all of that in Python with one function call, from keywords, product links or ASINs, on 20 Amazon stores, with no Amazon account or API key.

Quick start

# pip install apify-client
from apify_client import ApifyClient

client = ApifyClient("YOUR_APIFY_TOKEN")

run = client.actor("automly/amazon-products-scraper---fast-efficient-with-sales-data").call(
    run_input={"keywords": ["garlic press"], "maxResultsPerKeyword": 20}
)

for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    bsr = item["best_sellers_rank"][0]["rank"] if item.get("best_sellers_rank") else None
    print(item["asin"], item["price"], bsr, item.get("estimated_monthly_sales"))

That prints the ASIN, price, main-category rank and estimated monthly sales for the top 20 garlic presses on amazon.com. The rest of this post covers the inputs, the output, cost and limits.

Why not the official Amazon API?

Amazon has two developer APIs, and both come with conditions. The Product Advertising API (PA-API 5.0) that most old tutorials use is gone: Amazon deprecated it, and calls to it now return 403 Forbidden. Its replacement is the Creators API.

Option	Access needed	Cost	What you get
Creators API (replaces PA-API)	Amazon Associates account with at least 10 qualifying sales in the past 30 days. Access is lost after 30 straight days without referred sales (source)	No fee listed. Quota starts at 1 request/second and 8,640/day, then grows with the sales you refer (source)	Search, item details, variations, categories, offers and sales rank. No reviews or ratings resource (resources)
Selling Partner API	A Professional seller account (Individual accounts can't build private apps) plus developer registration	Professional plan is $39.99/month. Amazon dropped the developer fees it had planned for 2026	Catalog details and sales ranks (Catalog Items API), but its use policy limits it to helping sellers run their Amazon business
This scraper	Free Apify account	$5.55 per 1,000 results	Price, rating, reviews count, BSR per category, "bought in past month", monthly sales and revenue estimates

So the official route works if you're an active affiliate who links back to Amazon, or a seller managing your own business. For product research or price monitoring across any product, it's a poor fit. I also couldn't find Amazon's "bought in past month" figure or any sales estimate in either API's docs.

Step 1: Install the client and get a token

pip install apify-client

Sign up at apify.com, then copy your API token from Settings > API & Integrations. Keep it in an environment variable rather than in your code.

Step 2: Scrape Amazon search results by keyword

keywords takes a list. Each keyword returns up to maxResultsPerKeyword products (default 50, max 500). marketplace picks the store and defaults to com.

import os
from apify_client import ApifyClient

client = ApifyClient(os.environ["APIFY_TOKEN"])
ACTOR = "automly/amazon-products-scraper---fast-efficient-with-sales-data"

run = client.actor(ACTOR).call(run_input={
    "keywords": ["garlic press", "yoga mat"],
    "maxResultsPerKeyword": 100,
    "marketplace": "com",
})

Each row keeps search_keyword, search_position and a sponsored flag, so you can see where a product sits in the results as well as what it is.

Step 3: Scrape Amazon prices by ASIN

If you already have product IDs, pass them in asins. Product links go in startUrls. You can mix all three inputs in one run.

run = client.actor(ACTOR).call(run_input={
    "asins": ["B0DJ28XG7Z", "B00HEZ888K", "B098WNQ44M"],
    "startUrls": [{"url": "https://www.amazon.com/dp/B09QBDRCRK"}],
    "marketplace": "com",
})

for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(item["asin"], item["price"], item["list_price"], item["discount_percent"])

Run that on a schedule and you have a basic Amazon price tracker.

Step 4: Read the output

One row per product. Here's a real one, shortened:

{
  "asin": "B0DJ28XG7Z",
  "url": "https://www.amazon.com/dp/B0DJ28XG7Z",
  "title": "Kitchen Garlic Press, Rust Proof & Dishwasher Safe Garlic Mincer...",
  "brand": "SRISE",
  "price": "$6.99",
  "list_price": "$7.99",
  "discount_percent": 13,
  "rating": 4.6,
  "review_count": 1478,
  "best_sellers_rank": [
    {"rank": 886, "category": "Kitchen & Dining"},
    {"rank": 3, "category": "Garlic Presses"}
  ],
  "bought_in_past_month": "3K+",
  "estimated_monthly_sales": 3149,
  "estimated_monthly_revenue": "$22,011.51",
  "availability": "In Stock",
  "is_prime": true,
  "badges": ["Amazon's Choice"],
  "search_keyword": "garlic press",
  "search_position": 2
}

The full row also has the star breakdown, category path, bullet points, description, specs, variants with their own ASINs, images, coupon, seller, related ASINs and a scraped_at timestamp.

Step 5: Save to CSV

import csv

fields = ["asin", "title", "price", "rating", "review_count",
          "bought_in_past_month", "estimated_monthly_sales", "estimated_monthly_revenue"]

with open("amazon_products.csv", "w", newline="", encoding="utf-8") as f:
    writer = csv.DictWriter(f, fieldnames=fields + ["bsr"], extrasaction="ignore")
    writer.writeheader()
    for item in client.dataset(run["defaultDatasetId"]).iterate_items():
        top = item["best_sellers_rank"][0] if item.get("best_sellers_rank") else {}
        writer.writerow({**item, "bsr": top.get("rank")})

If you prefer pandas, pd.DataFrame(client.dataset(run["defaultDatasetId"]).list_items().items) does the same in one line.

How far to trust the sales estimate

The monthly sales number is derived from the product's Best Sellers Rank, and revenue is units times the current price. I use it to compare products against each other and to size a niche, and that's where it's good. For a single product, real sales can be about 55% above or below it, so don't treat it as an exact figure.

Estimates exist on amazon.com, .co.uk, .de, .fr, .it, .es, .ca, .com.mx and .in, for most main categories. On the other stores you get every other field, just not the estimate. Products with no BSR, like brand-new listings, have none either.

What else can you scrape?

The same code works on 20 Amazon stores: set marketplace to co.uk, de, fr, it, es, ca, com.mx, in, com.br, co.jp, com.au, ae, sa, nl, se, pl, com.tr, sg or eg. Prices and text come back in that store's language and currency.

For other jobs there are three related scrapers: Amazon Search Scraper for keyword rank tracking, Amazon Best Sellers Scraper for the top 100 of any category, and Amazon Reviews Scraper for review text. The first two give you ASINs you can feed straight into Step 3.

Scaling

Raise maxResultsPerKeyword up to 500 and add more keywords.
Set "fullDetails": false when you only need what the search list shows (ASIN, title, price, rating, image, sponsored flag). It's much faster, but you lose BSR and the sales estimates.
Want every variant priced? Each row lists its variants' ASINs; pass them back in through asins.
For daily tracking, save your input as a task in Apify and attach a schedule. Set a maximum cost per run if you want a hard budget.

No-code option: the Apify Console

You don't have to write Python at all. Open the actor page, click Try for free, type your keywords or ASINs, pick a marketplace and click Start. When it finishes, download the results as CSV, Excel or JSON, or send them to Google Sheets.

What it costs

You pay $5.55 per 1,000 results on every Apify plan, plus a start fee of $0.05 per GB of memory (that's $0.10 per run at the default 2 GB) and the run's platform usage. Every Apify account gets $5 of free usage each month, which covers several hundred products. That's enough to test a niche or two before paying anything.

Responsible use

Stick to public product data and follow Amazon's terms. Some fields, like seller names, can count as personal data under GDPR or CCPA, so have a legitimate reason to collect them and check with a lawyer if you're unsure.

Links

If you want Amazon prices, ranks and sales estimates without building and maintaining your own scraper, try Amazon Scraper - Product Prices, BSR & Sales Estimates on Apify. If a field you need is missing, request it on the actor's Issues tab.