Facebook Search Scraper - Public Posts by Keyword

Search public Facebook posts by keyword and get one row per post: the text, who posted it and in which group, the link, the exact time, the media and the reaction, comment, share and view counts.

Type a word or phrase and the actor returns the public posts Facebook's own search finds for it, without signing in. The catch: Facebook ends one search after about 65 to 100 posts. Give it a date window and it goes through the dates a week or so at a time, which is how you get hundreds of posts for a busy topic.

Input	Search terms, plus optional dates, recent posts or videos only
Output	One row per post: text, author, group, link, time, media type, image and video links, reactions, comments, shares, video views
Ceiling	About 65 to 100 posts per term for a plain search; up to 1,000 per term with a date window; 100 terms a run
Speed	About 160 to 190 posts a minute in test runs
Account needed	None from you
Price	$0.46 per 1,000 posts, flat on every plan. The free plan's $5 a month covers about 10,800 posts

🔍 What Facebook Search Scraper does

It searches Facebook's public posts the way the search box does for someone who isn't signed in: posts from Pages, public groups and public profiles. Each search term is its own search.

Every post is checked before it goes into your dataset:

• The words. A post is kept when every word you searched appears in its text, its author's name or its group's name. Facebook also returns looser matches, and those are left out unless you switch the check off.
• The dates. Facebook applies a date window loosely, and on some searches its Recent posts filter mixes in old posts. Each post's time is checked: anything outside your dates, or older than a week with recent posts, is dropped.
• Videos. With videos only, anything that isn't a video post is dropped.

When Facebook has nothing that matches, it still returns posts, just unrelated ones: a search for gibberish gets about ten. So if none of the first page's posts contain your words, the search stops there with a free note.

📋 What data you get from each Facebook post

What you get	Field
The post's id and link	postId, url
When it was posted, in UTC and as a Unix time	time, timestamp
The text	text
Who posted it	authorName, authorId, authorUrl
The public group it was posted in, if any	groupName, groupId, groupUrl
What it carries	mediaType (photo, album, video, link, text, shared post), isReel
Images, largest first	imageUrls
The video's page, cover and length	videoUrl, videoThumbnailUrl, videoDurationSeconds
The post a share points to	sharedPostUrl
Reactions, comments, shares, video views	reactionsCount, commentsCount, sharesCount, videoViewsCount
Which search found it, and its place in that search	searchTerm, position
Whether every searched word is in it	containsSearchWords

▶️ How to scrape Facebook posts by keyword

1. Open Facebook Search Scraper and click Try for free.
2. Type one or more words or phrases into Search terms, one per line.
3. Set Posts per search term. For more than about 100 posts on one term, fill in Posted from.
4. Tick Recent posts or Videos only if you want them.
5. Click Start, then download the rows as JSON, CSV or Excel, or read them through the Apify API.

Start it with every field empty to see the row shape first: one labelled sample post, not charged.

💰 How much does it cost to scrape Facebook posts?

$0.46 per 1,000 posts. Flat on every Apify plan, no volume tiers. On the free plan, the $5 Apify gives you each month covers about 10,800 posts.

You pay for posts that come back. The sample row and every note row are free, and so are posts the checks left out (missing words, outside your dates, not a video) and posts that turn up twice. If you set a maximum charge for the run, it stops before the first post it couldn't pay for.

📥 What you give it

{
  "searchTerms": ["electric car", "coffee"],
  "maxPostsPerSearch": 300,
  "startDate": "2026-08-01",
  "endDate": "2026-09-30"
}

Field	Default	What it is
searchTerms	none, the box starts at electric car	Words or phrases to search for, one per line, in any language. Up to 100.
maxPostsPerSearch	50	The most posts from one term, 1 to 1,000. A plain search ends after about 65 to 100.
recentPosts	false	Facebook's Recent posts filter, checked: only posts from the last 7 days are kept. Can't be combined with dates.
startDate	none	Only posts from this day on, as YYYY-MM-DD, read in UTC.
endDate	none	Only posts up to and including this day. With only a start date, the window runs to today.
videosOnly	false	Only video posts, reels included.
onlyPostsWithEveryWord	true	Keep only posts where every searched word appears in the text, the author's name or the group's name. Off returns everything Facebook's search does, marked with containsSearchWords.

Dates have to be YYYY-MM-DD. Anything else, like "last week", stops the run before it searches, with a free note saying what to write. So do a start date in the future, a start after the end, and recent posts together with dates.

📤 What you get back

A real row from a test run on 3 October 2026 (search "coffee"), image link shortened:

{
  "recordType": "post",
  "postId": "1083580981083414",
  "url": "https://www.facebook.com/firstbloomcoffee/posts/pfbid0AbckUBKoknuamg6NpvW9rjTVBHfTxBWjsg9jaLM2VV1GYdWch6sP8YNWzzoUV3j6l",
  "text": "👻 Announcing Cleveland’s first ever coffee crawl!\n\n🎃 Mark your calendars for the last week of October when all your favorite coffee shops in town will be featuring a special $5 drink!\n\nIt’s like trick-or-treating for grown-ups. 😜",
  "time": "2026-10-01T14:29:12.000Z",
  "timestamp": 1790864952,
  "authorName": "First Bloom Coffee and Bakery",
  "authorId": "100082945672195",
  "authorUrl": "https://www.facebook.com/firstbloomcoffee",
  "groupName": null,
  "groupId": null,
  "groupUrl": null,
  "mediaType": "photo",
  "isReel": false,
  "imageUrls": ["https://scontent-iad3-1.xx.fbcdn.net/v/t39.99422-6/828831765_...n.png?..."],
  "videoUrl": null,
  "videoThumbnailUrl": null,
  "videoDurationSeconds": null,
  "sharedPostUrl": null,
  "reactionsCount": 871,
  "commentsCount": 223,
  "sharesCount": 279,
  "videoViewsCount": null,
  "searchTerm": "coffee",
  "position": 5,
  "containsSearchWords": true,
  "scrapedAt": "2026-10-03T16:28:56.833Z"
}

Field	Worth knowing
authorUrl	null on about one post in seven, mostly people posting in groups; authorId is always there
authorName	"Anonymous participant" on anonymous group posts
text	null on the rare photo or video posted without a caption
groupName, groupId, groupUrl	null unless the post was made in a public group
imageUrls, videoThumbnailUrl	Facebook's own links; they stop working after a few days, so save the files you need
videoViewsCount	Only on video posts
reactionsCount	All reactions together, not split by type
position	The order this search returned the post in, after the checks

🧾 Reading the output

Every real post has recordType: "post". Filter on that and you have only the posts you paid for.

An empty run gives one sample row with _sample: true and recordType: "sample". Notes have _diagnostic: true, ok: false, an errorCode and a plain message, and are never charged:

errorCode	Means
NO_MATCHES	Facebook returned posts, but none of the first ones contain your words, so the search stopped
NO_RESULTS	Nothing to return: no posts, or none inside your dates, none from the last week with recent posts, or none that were videos
BAD_INPUT	A setting or term couldn't be used; the message says what to change
NOT_ANSWERED	Facebook didn't answer that search this time; try it again later
PARTIAL	Facebook stopped answering partway through a term; the posts before that are in the dataset
NOT_REACHED	The run stopped (time, your maximum charge, or Facebook not answering) before this term
SEARCH_UNAVAILABLE	Facebook's post search didn't accept the request at all; nothing was searched
MAX_CHARGE_TOO_LOW	Your maximum charge doesn't cover a single post

The key-value store also holds RUN_REPORT: each term's status, posts returned, pages read and what the checks left out.

💡 What people use it for

• Hearing what people say about a brand or product. A shop owner searching their own name finds the group posts nobody tagged them in.
• Topic research. What people post about electric cars, a diet or a new phone, with the counts showing which posts landed.
• Daily monitoring. Recent posts on a schedule, keeping only post ids you haven't seen.
• Finding the groups where a topic lives. groupName across a few hundred posts shows which public groups talk about it most.

A recipe with our other Facebook actors: search a topic here, keep the posts with the most commentsCount, then pass their url values to the Facebook Comments Scraper to read the conversation under each.

🚧 What it does not do

• No private posts, private groups or friends-only posts. Only what Facebook shows people who aren't signed in.
• A plain search ends after about 65 to 100 posts, however many exist. Dates are the way past that.
• Results lean towards what a visitor in the United States sees, so English words bring mostly American posts.
• No photos-only filter. Facebook doesn't show photo results to people who aren't signed in.
• With dates or recent posts, videos only keeps the video posts among the normal results, so it reads more pages for fewer posts.
• No comments, only their count. Those come from the Facebook Comments Scraper.
• No follower counts or contact details for the authors. Those come from the Facebook Pages Scraper.

🧭 Which Facebook scraper do you need?

If you want	Use
Public posts about a word or phrase, from anywhere on Facebook	This one
Every recent post of Pages you already know	Facebook Posts Scraper
The posts inside a public group	Facebook Groups Scraper
The comments under posts or reels you found	Facebook Comments Scraper
Name, likes and followers of the Pages behind the posts	Facebook Pages Scraper
Events about a topic or in a city	Facebook Events Scraper
The video file of a reel or video post	Facebook Video Downloader

❓ Questions people ask

Do I need a Facebook account to search Facebook posts?

No. It reads the public posts Facebook shows people who aren't signed in. Nothing of yours is used and nothing is posted.

Why are the posts different from what I see when I search on Facebook?

Your own search is shaped by your friends, groups, language and country. A run sees what Facebook shows an anonymous visitor, leaning towards one in the United States. A private window won't let you check, because Facebook doesn't open its search page there, but every post link in the dataset does open in one.

How do I get more than about 100 posts for one search term?

Fill in Posted from, and Posted until if you want an end, then raise Posts per search term. The window is searched in slices, newest first, and each slice brings its own posts. In a test, "pizza" from January to September 2026 reached the 1,000-post cap within the window's last 19 days, every post inside the dates.

Why did my search return a note instead of posts?

NO_MATCHES means Facebook's results didn't contain your words, which is what it does when it has nothing on a topic. NO_RESULTS means nothing was left after your dates, the recent check or the videos filter. Neither is charged.

Can I call it from code or connect it to an AI assistant?

Yes. The API tab has code for Python, JavaScript and the command line. For Claude, ChatGPT or another MCP client, connect https://mcp.apify.com/?tools=fetch-actor-details,dami_studio/facebook-search-scraper. Either way it runs on your Apify account at the same price.

Is scraping Facebook posts legal?

It reads public posts only. Names and posts are still personal data under laws such as GDPR, and Facebook has its own terms, so have a reason and read them. Apify's write-up on scraping and the law is a starting point; we are not lawyers.

🆘 If something breaks

Open an issue on the Issues tab with the run link and the search terms you used. A note row's errorCode and message help too.