X (Twitter) Boolean Search Scraper | Posts & Trends

Find public X posts with one precise Boolean query. Combine keywords, hashtags, exact phrases, AND, OR, NOT, parentheses, dates, language, media type, engagement thresholds, and author filters without managing cookies or a developer account.

The Actor validates your expression, creates a safe search plan, discovers candidate posts, checks the complete Boolean logic locally with Unicode-aware whole-term matching, removes duplicates, and streams matching rows to the dataset.

What You Can Do

• Monitor brands, products, executives, competitors, or events.
• Find purchase intent, complaints, feature requests, and campaign conversations.
• Compare alternative names, spellings, hashtags, and phrases in one run.
• Exclude job posts, giveaways, spam terms, or unrelated meanings.
• Keep only recent, popular, verified, image, video, reply, or original posts.
• Export structured results to JSON, CSV, Excel, XML, RSS, or the Apify API.

Boolean Search

Supported syntax:

Operator	Meaning	Example
AND	Both expressions must match	OpenAI AND agents
OR	Either expression may match	OpenAI OR Anthropic
NOT	Exclude a matching expression	AI AND NOT hiring
"..."	Exact phrase	"artificial intelligence"
( ... )	Control precedence	(OpenAI OR Anthropic) AND agents
#hashtag	Match an exact hashtag	#AI OR #MachineLearning

Terms use complete Unicode word boundaries. A query for IA matches the standalone term IA, but not tia, viaje, or diario. Matching is case-insensitive and accent-insensitive.

Validate and explain an expression before a paid run with the free FlujoDato Boolean Search Validator.

Input Example

{
  "searchQuery": "(OpenAI OR Anthropic) AND (agents OR automation) AND NOT hiring",
  "resultsLimit": 100,
  "searchSection": "latest",
  "searchCoverage": "efficient",
  "contentType": "all",
  "language": "en",
  "oldestPostDate": "",
  "newestPostDate": "",
  "minimumViews": 0,
  "minimumLikes": 5,
  "minimumReposts": 0,
  "minimumReplies": 0,
  "minimumQuotes": 0,
  "minimumBookmarks": 0,
  "verifiedAuthorsOnly": false,
  "excludeReplies": false,
  "excludeReposts": true,
  "strictBooleanSyntax": false
}

Input Fields

Field	Description	Default
searchQuery	Keyword, hashtag, phrase, or Boolean expression	Example query
resultsLimit	Maximum unique matching posts saved	100
searchSection	latest or top ranking	latest
searchCoverage	efficient scans up to 3x; comprehensive scans up to 6x and adds exact-term and hashtag variants that preserve the Boolean condition	efficient
contentType	All, text, media, images, or videos	all
language	All languages or one supported language	all
oldestPostDate	Inclusive start date	Empty
newestPostDate	Inclusive end date	Empty
minimumViews	Minimum reported views	0
minimumLikes	Minimum likes	0
minimumReposts	Minimum reposts	0
minimumReplies	Minimum replies	0
minimumQuotes	Minimum quote posts	0
minimumBookmarks	Minimum reported bookmarks	0
verifiedAuthorsOnly	Keep verified authors only	false
excludeReplies	Remove replies	false
excludeReposts	Remove reposts	false
strictBooleanSyntax	Require explicit operators	false

Output Example

{
  "query": "OpenAI AND AI",
  "normalizedQuery": "OpenAI AND AI",
  "matchedTerms": ["OpenAI", "AI"],
  "discoveryQuery": "OpenAI AI",
  "searchSection": "latest",
  "tweetId": "1234567890123456789",
  "url": "https://x.com/example/status/1234567890123456789",
  "text": "A new OpenAI release for practical AI agents.",
  "hashtags": ["#AI"],
  "publishedAt": "2026-09-07T12:00:00Z",
  "language": "en",
  "contentType": "text",
  "mediaUrls": [],
  "videoUrls": [],
  "likeCount": 42,
  "repostCount": 7,
  "replyCount": 3,
  "quoteCount": 1,
  "viewCount": 2500,
  "bookmarkCount": 4,
  "isReply": false,
  "isRepost": false,
  "isQuote": false,
  "authorUsername": "example",
  "authorName": "Example",
  "authorProfileUrl": "https://x.com/example",
  "authorVerified": true,
  "authorFollowerCount": 12500,
  "scrapedAt": "2026-09-07T12:01:00Z"
}

Every row represents one unique matching post. Numeric zero values remain 0. Optional unavailable values are returned as null or empty arrays according to their type.

Pricing And Search Coverage

resultsLimit is a maximum, not a guarantee. Search feeds are ranked, can repeat posts across pages, and can end before the requested count.

• Efficient can inspect up to 3 times the requested result limit.
• Comprehensive can inspect up to 6 times the requested result limit for restrictive queries.
• When an end date is set, Efficient can use up to 0.5 successful pages per requested result and Comprehensive up to 1 page per result, shared across the complete search plan.
• Restrictive AND, NOT, date, media, author, or engagement filters may return fewer results.
• A successful page containing new candidate posts is charged even if local filters reject every candidate.
• Failed requests, retries, empty pages, and pages containing only previously seen posts are not charged.
• Each saved dataset item is charged once. Duplicate posts are never saved or charged as results.

The start date is sent to the upstream search, while the end date is verified locally. Boolean-qualified posts rejected only by the end date or another local filter keep that source active so historical pagination is not stopped as irrelevant.

Before discovery begins, the Actor estimates the maximum result and page charges required by the plan. If the run can afford at least one successful page and one result, it can start in partial mode, stream affordable output, and stop safely when the charge limit is reached. A run is rejected before provider use only when it cannot fund that minimum useful operation.

Reliability

• Results are streamed after every completed page, so partial output survives an interrupted run.
• Stable post IDs prevent duplicate result billing.
• Temporary network, rate-limit, and server failures retry after 4, 16, and 64 seconds.
• Repeated cursors and duplicate-only pagination cycles stop automatically.
• Live progress is available in RUN_PROGRESS; the final counters are stored in RUN_SUMMARY.
• Credentials, endpoints, provider payloads, and pagination tokens are never written to the dataset or logs.

Tips For Better Results

1. Start with 25 to 100 results in Efficient mode.
2. Test each positive term separately if a complex query returns few matches.
3. Add engagement filters gradually; high thresholds can remove most candidates.
4. Use Latest for monitoring and Top for research on prominent conversations.
5. Switch to Comprehensive only when the first sample confirms that the query is useful.

Responsible Use

Use only public data and follow applicable laws, privacy requirements, platform terms, and your organization's data retention policies. Do not use the Actor for harassment, surveillance, discrimination, or collection of sensitive personal data.

Support

For reproducible support, share the Actor run ID, input with secrets removed, expected behavior, and the relevant log lines. Do not send API keys or private credentials.