Twitter User Scraper: profiles, followers and following

Which of the seven charges will your run fire?

Naming an account costs $0.004. Asking about its network costs $0.016 a time, and three switches decide how many of those fire. Two of the three are on before you touch anything.

• twitterHandles · startUrls · twitterUserIds

  $0.004per account named

  No users included

  Three ways to name an account. Identical price, so pick whichever matches your data.

  ["elonmusk"] or 44196397

• On before you touch anything

  getFollowers · getFollowing

  +$0.032per profile, unasked

  ~70 users each included

  Both prefill to true in the schema, so a plain handle lookup already fires two network queries you did not ask for.

  getFollowers: false

• getRetweeters

  $0.016per tweet

  ~70 retweeters included

  Also prefills true. Applies to tweet URLs in startUrls, and returns who retweeted them.

  x.com/NASA/status/198134…

• searchTerms

  $0.008per search

  ~20 users included

  Finds accounts by keyword instead of naming them. Half the price of a profile query and it brings twenty users with it.

  ["climate scientist"]

Every user past a query’s included page costs $0.0004, and getAbout adds $0.0002 a profile for creation date, username history and verification timestamps. maxItems is the ceiling on all of it.

Three questions about people

Look an account up, pull the network around it, or find accounts you cannot name yet. The three cost quite differently, which is the whole reason to know which one you are doing.

{
  "twitterHandles": ["NASA", "ESA", "JAXA_en", "isro", "csagc"],
  "getFollowers": false,
  "getFollowing": false,
  "getRetweeters": false,
  "maxItems": 5
}

{
  "type": "user",
  "userName": "exampleUser",
  "id": "1467931973616386052",
  "name": "Example User",
  "description": "Bio description with links and hashtags",
  "location": "San Francisco, CA",
  "followers": 1147899,
  "following": 561,
  "isVerified": true,
  "isBlueVerified": true,
  "verifiedType": "business",
  "createdAt": "Mon Dec 06 19:00:41 +0000 2021",
  "statusesCount": 2104
}

{
  "twitterHandles": ["NASA"],
  "getFollowers": true,
  "getFollowing": false,
  "maxItems": 500
}

{
  "followerOf": "elonmusk",
  "type": "user",
  "userName": "exampleUser",
  "followers": 1147899,
  "following": 561,
  "protected": false,
  "canDm": true,
  "listedCount": 1294,
  "mediaCount": 139
}

{
  "searchTerms": ["climate scientist", "oceanographer"],
  "getFollowers": false,
  "getFollowing": false,
  "maxItems": 200
}

{
  "type": "user",
  "userName": "example_researcher",
  "name": "Example Researcher",
  "description": "Ocean carbon. Views my own.",
  "followers": 18422,
  "isBlueVerified": false,
  "listedCount": 204,
  "favouritesCount": 1450
}

What an account record holds

One shape whatever produced it, and the listing publishes an availability guarantee for each field: eleven are always present, two depend on whether the account filled them in.

• 6

  Identity

  Always present

  • id, userName
  • name
  • url, twitterUrl
  • profilePicture
  • description

• 5

  Size and reach

  Always present

  • followers, following
  • statusesCount
  • listedCount
  • mediaCount
  • favouritesCount

• 4

  Status

  Verification kept separate

  • isVerified, legacy
  • isBlueVerified, X Premium
  • verifiedType
  • protected, canDm

• 3

  Context

  Some conditional

  • createdAt
  • location, if the user set one
  • coverPicture, if set
  • followerOf, on network results

isVerified and isBlueVerified are not the same signal and the listing keeps them apart on purpose: one is legacy verification, the other is a paid subscription. Filtering on the wrong one is the usual way an influencer shortlist goes wrong.

What will surprise you

Five things, and the first two are the reason a first run costs more than expected.

• 01

  Two network switches are on by default

  getFollowers and getFollowing prefill to true in the input schema, and getRetweeters does too. The README’s table says all three are false. The schema is what the form loads, so assume on and set them explicitly.

• 02

  A run needs at least five objects

  The actor states a minimum of 5 objects per run to cover its operating cost. For one profile with no network, the documented workaround is to list the handle five times.

• 03

  Each direction of a network is its own query

  Followers and following are $0.016 each, with their own included page. Retweeters is a third. Nothing bundles them.

• 04

  Included pages are approximate

  Roughly 70 users on a network query, roughly 20 on a keyword search. They are first-page sizes, not guarantees, so budget on the item rate rather than the free tier.

• 05

  Suspended accounts are excluded unless you ask

  includeUnavailableUsers is off, so suspended and unavailable accounts are dropped. Turn it on when a gap in a follower list would itself be the finding.

Free Apify accounts run it in demo mode: 5 runs a month, capped at 10 items.

What each kind of run costs

A run costs the fee for every query it fires plus $0.0004 for each user past the included pages. The rows below are the listing’s own worked examples, and all seven reproduce from the published event prices.

The estimator above prices one kind of query at a time, because that is what it can do honestly: a run that mixes a profile lookup with a followers pull adds both fees, as rows four to six show. Computed from the seven charge events the Apify Store API reports; the listing rounds its retweeters example to $0.07 where the arithmetic gives $0.068.

Who scrapes X accounts rather than tweets?

Five kinds of work need the people rather than the posts: influencer research, lead generation, audience overlap, account intelligence and academic network analysis.

• Influencer research

  Find accounts, then look them up

  Search a niche, then pull full profiles on the shortlist and filter on isBlueVerified and listedCount rather than follower count alone

  $0.04 per 100 found

• Lead generation

  Pull a network, followers only

  Export the followers of an account your buyers already follow, with bios and locations for qualification

  $0.19 per 500 followers

• Audience overlap

  Pull a network on two accounts

  Take both follower lists and intersect them to see how much of a competitor’s audience you already share

  $0.78 per 1,000 each way

• Account intelligence

  Look up accounts, with getAbout

  Check creation dates, username history and verification timestamps when an account’s age is the question

  $0.0042 per profile

• Network researchers

  Pull a network, at scale

  Build follower graphs for community detection, where followerOf is the edge and the record is the node

  $0.0004 per edge

Wiring it into a pipeline

The actor is apidojo/twitter-user-scraper. The one thing to encode deliberately is the switches: leave them out of your run input and you inherit the schema defaults, which fire two network queries per account.

from apify_client import ApifyClient

client = ApifyClient("YOUR_APIFY_TOKEN")

# Always state the switches. Omitted, they inherit the schema's true and
# each account quietly fires two $0.016 network queries.
run = client.actor("apidojo/twitter-user-scraper").call(run_input={
    "twitterHandles": ["NASA", "ESA", "JAXA_en", "isro", "csagc"],
    "getFollowers": False,
    "getFollowing": False,
    "getRetweeters": False,
    "maxItems": 5,
})

for u in client.dataset(run.default_dataset_id).iterate_items():
    # Legacy verification and paid Premium are different fields
    print(u["userName"], u["followers"], u["isVerified"], u["isBlueVerified"])

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: process.env.APIFY_TOKEN });

// Audience overlap: two follower lists, intersected.
const run = await client.actor('apidojo/twitter-user-scraper').call({
  twitterHandles: ['NASA', 'ESA'],
  getFollowers: true,
  getFollowing: false,
  getRetweeters: false,
  maxItems: 4000,
});

const { items } = await client.dataset(run.defaultDatasetId).listItems();

// followerOf is the edge: which account's network produced this record.
const sets = Object.groupBy(items, (u) => u.followerOf);
const [a, b] = Object.values(sets).map((g) => new Set(g.map((u) => u.id)));
console.log([...a].filter((id) => b.has(id)).length, 'shared followers');

curl -X POST \
  "https://api.apify.com/v2/acts/apidojo~twitter-user-scraper/run-sync-get-dataset-items?token=YOUR_APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"searchTerms":["climate scientist"],"getFollowers":false,"getFollowing":false,"maxItems":200}'

Datasets export as CSV or stream over the API. Every key above is in the actor’s input schema, and the switches are named in full precisely because leaving them out is the expensive option.

Twitter User Scraper FAQ

Where Twitter User Scraper fits