Most research starts with twenty open tabs. Other companies' websites, their prices, their reviews, their posts. You read, take notes, forget half of it. Apify pulls the same thing as data, and Claude reads it for you.
Apify is a platform with over 85,000 ready-made tools that pull public data from the web. Instagram, TikTok, YouTube, Google Maps, Google searches, webshops, reviews, ads and ordinary websites. Apify calls the tools Actors. You pick one, paste a link and get the numbers back as a table. Connected to Claude, you ask in plain language instead. Claude picks the tool, runs it and analyses the result for you.
The rule is simple: anything you can see in a browser, Apify can pull. What the three biggest players in your field write on their about page. What a photographer in Aarhus charges for a portrait. What customers complain about at the supplier you are considering. Questions that used to take an afternoon, Claude answers while you drink your coffee.
What you can use Apify for
Apify lists over 30 use cases on its own website. These eight are the ones that make sense for a small business.
1. How other companies do it
Website Content Crawler pulls the text from entire websites. Services, prices, how the offers are built, what they promise, and what they leave out. Twenty companies' sites in one file, which Claude reads across and compares. For a single page I use Firecrawl. For twenty at once, Apify.
2. Competitors on social media
Instagram Scraper pulls posts, reels, captions, likes, comments and views from public profiles. TikTok and YouTube each have their own tool. That is the use my own example further down is about.
3. Prices and products in webshops
The e-commerce tools track prices, stock and new products at the shops you want to keep an eye on. You see a price change the day it happens, instead of discovering it when a customer mentions it.
4. Reviews, as customers write them
From Google, Trustpilot, Booking.com and webshops. Your own, so you see what customers praise and miss. A whole industry's, so you know what customers are tired of before you write your next page.
5. Ads other people pay for
The tools for Meta's Ad Library pull a page's active ads: copy, images and how long each ad has been running. If an ad runs for three months, it makes money. It is the cheapest market research there is.
6. Markets and leads on Google Maps
Google Maps Scraper pulls out businesses by keyword and area, with address, phone, website and number of reviews. How many yoga studios are there in Køge, and which of them have fewer than ten reviews? For you if you sell to other businesses, or are thinking about opening somewhere.
7. Hashtags and trends
Which hashtags are used together on Instagram, and which topics are moving on TikTok right now. Data instead of guesses when you pick a topic.
8. Content for your own AI
The same Website Content Crawler, turned inward. Your own website as clean files, so Claude knows everything you have written. Or a supplier's documentation, so you can ask instead of searching.
How I use it myself: the competitor analysis
My standing run is an analysis of the accounts I share a field with on Instagram. The list has eight accounts. One direct Danish competitor, six larger accounts in the same field, most of them international, and Claude's own Instagram as a news source. Every account is on the list with one line about why it is there. Otherwise the list drifts into randomness.
Once a month I ask Claude to pull the latest 15 posts from each account, only the last four weeks. Then it sorts by likes and comments, and I look for four things.
Which format carries
The split between reels and carousels among the ones that are growing. Not among everyone.
Which CTA they use
Comment keyword, follow for part 2, save this. If a post jumps from 20 to 200 comments, there is a mechanic behind it, not luck.
Which hook type keeps coming back
In their best posts. List, admission, opinion, POV. The mechanic can be translated into my world. The words cannot.
What they leave alone
The gap is worth more than the copy. If everyone writes how-to, there is open ground on what does not work, and on the craft itself.
Old viral posts do not count. They distort the picture and make you chase something that worked last year.
And before I conclude anything, I check Meta's Ad Library. When I ran the analysis the first time, it turned out that the most viewed reel from my Danish competitor had been running as an ad for two and a half months. The follower growth was bought. That changed the lesson. The difference was in the funnel behind the posts.
The conclusion goes into every content proposal as one line: why now. With a number from the analysis, a gap in the market or a piece of news. A proposal without that line is a hunch.
The line I do not cross: mechanics, format and choice of topic may inspire. Wording, sentences and angles may not. Last check, every time: could this sentence stand word for word at the account I just looked at? Then I rewrite.
How to get started
1. Create an account with Apify
Through the link at the bottom. The free plan gives you 5 dollars a month to pull data for, no card. Instagram Scraper costs 2.70 dollars per 1,000 results on the free plan, so the first analyses stay within the free allowance.
2. Connect Apify to Claude
In Claude: Customize, then Connectors, plus, Add custom connector. Name: Apify. URL: https://mcp.apify.com. Add, and approve in the browser. You log in to Apify once, no key to copy. On Claude's free plan there is room for one custom connector. Switch it on in the conversation under the plus, Connectors. Each call to an Actor asks permission the first time: Allow once or Always allow.
3. Write your question down before you pull anything
One sentence: "What do the ten biggest in my field charge per hour?" or "What is working for my competitors right now?" The question decides which tool Claude picks, and how little needs to be pulled. For the competitor analysis the answer is a list: five to eight accounts. One or two direct competitors in the same language and market. Three or four larger accounts in the same field, preferably international, because they show mechanics before they arrive here. And one news source, for example the maker's own profile if you work with a particular tool.
4. Ask Claude to pull and sort
The prompt for the competitor analysis is below. Swap it for your own question when it is something else you want to know. Paste the full profile links, not usernames. The username field fails and returns nothing.
5. Check the Ad Library when it is about growth
Go to facebook.com/ads/library and search for the account's name. If their best reel is there as an ad, the growth is paid. Then the lesson is a different one.
Pull the latest 15 posts from these profiles with Apify's Instagram Scraper, only the last four weeks: [full profile links, one per line] Sort them by likes and comments. For the top ten, show me account, format, CTA and hook type in one table. Then tell me which format carries among the accounts that are growing, which CTA gets the most comments, and what none of the accounts write about.
Only pull the fields you need. The result from Apify is large, and Claude reads better when it gets username, type, caption, likes, comments, views and timestamp, and not everything.
Price
Free plan with 5 dollars a month. Starter costs 19 dollars a month and gives you 19 dollars to pull for. Eight accounts with 15 posts each is about 120 results. That is under half a dollar per run with Instagram Scraper. Every run costs, so run small, and only pull what you need.
I am an Apify affiliate, and the link on this page is an affiliate link. If you sign up through it, I get a share. The price is the same for you, and I have been running my analyses with Apify since before there was a link.
Apify is the eyes on the outside in my setup. The rest of the tools, and what each of them is hired to do, are in my tech stack.
