Scrape tweets by keyword without the official X API
Collect recent public X posts for a brand, phrase or hashtag and export text, dates, authors, URLs and visible engagement metrics.
Affiliate link. This workflow avoids official API setup; it does not remove source terms, privacy or copyright responsibilities.
Direct answer
How do you export tweets matching a keyword?
Set Twitter X Scraper to search, add one query per line in Input_Search, choose latest for recent posts and start with 50 results. Review relevance in the Dataset, then export selected columns to CSV or JSON.
You do not need to create an official X developer project for this Actor workflow.
- Mode
search- Sort
latestortop- Per-query limit
- 1 to 200 results
- Starting price
- From $4 / 1,000 results
Brand-monitoring input
Start with one phrase and 50 recent posts
{
"Input_Search": ["\"your brand\" problem"],
"TypeScraper": "search",
"filter": "latest",
"resultsLimit": 50
}- 01
Write a narrow question
Example: which product problems are mentioned next to the exact brand name this week?
- 02
Use quotation marks for names
"your brand" problemreduces matches where the words appear separately. - 03
Choose latest or top
Latest is better for monitoring. Top is useful for prominent posts but reflects X's relevance ranking.
- 04
Remove false matches
Review ambiguous brand names, reposts, spam, language and duplicate post URLs.
Query patterns
Change the search, not the whole workflow
| Question | Example query | What to review |
|---|---|---|
| Brand problems | "Acme" problem | Support requests and irrelevant name matches |
| Product comparison | "Acme" vs | Which alternatives are actually mentioned |
| Campaign hashtag | #AcmeLaunch | Organic posts versus promotional repetition |
| Purchase questions | "Acme" worth it | Context, date and geography |
Useful output
Keep evidence traceable
- Post URL or IDUse it as the unique key and source reference.
- Text and datePreserve enough context and normalize timestamps.
- Author handleTreat it as a public source identifier, not verified identity.
- Visible metricsRecord extraction time because likes and replies continue changing.
Budget control
Test each query for about $0.20
At the advertised starting rate of $4 per 1,000 delivered posts, 50 results estimate to $0.20.
| Queries | Results per query | Maximum results | Estimate from |
|---|---|---|---|
| 1 | 50 | 50 | $0.20 |
| 5 | 50 | 250 | $1.00 |
| 10 | 200 | 2,000 | $8.00 |
FAQ
Search decisions
Does this use the official X API?
No official developer project is required for the reviewed Actor. Apify handles the collection workflow.
Can I search multiple keywords?
Yes. Add queries separately and retain the source query in your analysis so results remain attributable.
Can I automate daily monitoring?
Yes. Save the input as a Task, schedule it and deduplicate by post ID. Use a frequency justified by the question.
First keyword run
Test one exact phrase with 50 recent posts
Measure relevance before adding more queries.