X / Twitter Scraper — Clean Data API

Extract public X (Twitter) posts into clean, duplicate-free JSON with engagement metrics, author profiles, and media.

Store promise: never fabricates posts to hit a quota. Empty or blocked sources → warnings + receipt, not filler rows.

No buyer API keys. Paste handles, URLs, search terms, lists, or conversation IDs — pay only for delivered unique posts.

Why this Actor

Need	What you get
Reliable post lookup	Public status URLs work with zero keys (public embed path)
Profiles, search, lists, threads	When the publisher enables a data transport (buyers still paste nothing)
Enriched fields	Likes, reposts, replies, quotes, views, bookmarks, author meta, media, reply/quote/retweet flags
Advanced search	Filters for users, language, verified, media, links, min engagement, date range
Clean pipelines	Canonical URLs, post-ID dedupe, exact maxResults (up to 50,000)
Transparent billing	PPE only for delivered unique posts; receipt is never billed

Quick start

Public post URLs (zero secrets)

{
  "postUrls": [
    "https://x.com/OpenAI/status/1950240351547248941"
  ],
  "maxResults": 20
}

Handles + search

{
  "handles": ["Apify"],
  "searchTerms": ["web scraping"],
  "searchType": "Latest",
  "language": "en",
  "excludeRetweets": true,
  "maxResults": 50,
  "includeReplies": false
}

Lists + conversations + filters

{
  "listUrls": ["https://x.com/i/lists/1234567890"],
  "conversationIds": ["1950240351547248941"],
  "fromUsers": ["OpenAI"],
  "minLikes": 10,
  "sinceDate": "2026-01-01",
  "maxResults": 100
}

Buyers never configure tokens. If timelines/search are unavailable, the run warns and still processes any post URLs it can resolve.

Output

record_type	Description	Billed?
post	One unique validated public post	Yes → validated-unique-post
receipt	Counts, transport, resource plan, warnings	Never

Example post (enriched)

{
  "record_type": "post",
  "id": "1950240351547248941",
  "url": "https://x.com/OpenAI/status/1950240351547248941",
  "text": "…",
  "author_handle": "OpenAI",
  "author_name": "OpenAI",
  "author": {
    "handle": "OpenAI",
    "name": "OpenAI",
    "is_blue_verified": true,
    "followers_count": 1000000,
    "profile_image_url": "https://…"
  },
  "published_at": "2025-07-30T12:00:00+00:00",
  "language": "en",
  "conversation_id": "1950240351547248941",
  "is_reply": false,
  "is_quote": false,
  "is_retweet": false,
  "like_count": 1200,
  "retweet_count": 340,
  "reply_count": 88,
  "quote_count": 12,
  "view_count": 50000,
  "bookmark_count": 90,
  "media": [],
  "source": "x_data_api"
}

source is one of: x_public_oembed | x_api_v2 | x_data_api (never names an upstream vendor).

Pricing (pay-per-event)

Event	When	Paid plans (≈)	Free plan (≈)
validated-unique-post	Each unique post delivered	$0.34–$0.40 / 1K	$8.00 / 1K (deterrent)

Plan	USD / post	USD / 1K
FREE	$0.00800	$8.00
BRONZE	$0.00040	$0.40
SILVER	$0.00039	$0.39
GOLD	$0.00037	$0.37
PLATINUM	$0.00035	$0.35
DIAMOND	$0.00034	$0.34

Not billed: receipts, duplicates, empty runs, warnings, failed fetches.

How keys work (publisher vs buyer)

Role	What they do
Buyer (Store user)	Input JSON only — no OpenAI, no X developer account, no third-party keys
Publisher	Sets a data transport once on the Actor so handles/search/lists work for everyone

This matches market leaders: “No API key” for the person running the Actor.

AI/sentiment is not bundled — export JSON and analyze in your own stack (or a separate analysis Actor later).

Input fields

Field	Description
postUrls	Public status URLs
handles	Profile timelines
searchTerms	Search queries (supports X advanced-search operators)
listUrls	Public list URLs
conversationIds	Root tweet IDs for thread/conversation pull
maxResults	Unique post cap (1–50,000)
includeReplies	Include replies on profile timelines
searchType	Latest or Top (third-party search)
fromUsers / mentionUsers	Author / mention filters
language	ISO 639-1 code (e.g. en)
onlyVerified / mediaOnly / linksOnly	Boolean search filters
excludeReplies / excludeRetweets	Drop replies / retweets from search
minLikes / minRetweets	Minimum engagement (third-party transport)
sinceDate / untilDate	YYYY-MM-DD date bounds
maxConcurrency	Parallel fetches (1–16)
resourceProfile	safe / speed / manual
memoryMbytes	Preferred run memory

Memory

Dynamic defaults typically 512–2048 MB (max 4096). Console override wins.

Limitations

• Public oEmbed path is for status URLs, not firehose search without publisher transport.
• Search windows depend on the publisher’s data transport (official X recent search is often ~7 days).
• minLikes / minRetweets apply on the third-party transport; official X API ignores them with a warning.
• Private/suspended posts fail with a warning — no synthetic substitutes.
• Comply with X Terms and applicable law for your use case.

Support

Include run ID + receipt when reporting issues. Target: first response within 24 hours when published.

Changelog

See ../CHANGELOG.md.