Threads Scraper - Posts, Reposts, Replies, Profiles & Search

Search public Meta Threads posts by keyword, collect an account's public posts, reposts or replies up to your selected limit, or find profiles by keyword. No login required. Export posts, media, public engagement metrics, and public profile data.

Before you start: Threads changes its public pages from time to time, and a change can briefly affect what any tool can collect until it is adjusted. Run a small sample first, check the output, then scale up. Results reflect what Threads shows publicly at the time of the run.

What You Can Do With It

• Brand & Mention Monitoring: watch brands, product terms, topics and campaigns as fresh public posts appear
• Competitor Analysis: compare public engagement across competitor accounts and posts
• Influencer Discovery: find creators by keyword or niche, with public follower counts
• Lead Generation: build account and prospect lists from public profile information
• Trend & Content Analysis: collect posts on a topic, hashtag, or campaign for downstream analysis
• Market Research: gather public posts on a product or category to read the conversation
• Date-bounded Research: narrow a search to a specific date range for time-sensitive work

Choose Your Workflow

Goal	Use this mode	What you enter
Collect posts from one or more accounts	user	Threads usernames or profile URLs
See what an account boosted, and when	user_reposts	Threads usernames or profile URLs
Read the replies an account wrote	user_replies	Threads usernames or profile URLs
Search public posts by topic	search	Keywords or hashtags
Discover account matches by keyword	profiles	Keywords, niches, or local phrases
Monitor fresh public mentions	search + recent	Brand, topic, or competitor keywords
Research a specific date range	search + date filters	Keyword plus start/end date

Ready-Made Examples

Each one is a pre-configured run you can start and then edit. Browse them all on the Examples tab.

Example	What you get
See which of your Threads posts perform best	Your posts with likes, replies, timestamps, and views when public, ready to sort
Find competitors' best-performing Threads posts	Several competitor accounts side by side in one export
Turn a Threads account into an AI knowledge base	Clean structured text with timestamps for a RAG pipeline
Audit a brand's most relevant Threads mentions	Brand mentions with public engagement fields and a keyword-relevance flag
Find creators posting about your niche	Accounts with followers, bio and profile links
See what people said about a product this week	The last 7 days of posts, newest first
Look up Threads accounts matching a keyword	An account shortlist with follower counts

How To Use

Input Configuration

Parameter	Type	Description	Example
mode	string	Scraping mode: user, user_reposts, user_replies, search, or profiles (the shorter reposts / replies also work)	"user"
usernames	array	Target usernames (for User Posts, Reposts and Replies modes, max 20). Accepts bare name, with @, or full Threads URL.	["zuck", "@nike", "https://threads.com/@meta"]
keywords	array	Keywords or hashtags (required in Search and Find Profiles, optional in the three user modes to tag matching rows, max 20)	["AI", "tech"]
search_filter	string	Search mode only. top shows Threads-ranked relevant results. recent prioritizes newest matching public posts.	"top"
start_date	string	Search mode only. Show posts from this date onward (YYYY-MM-DD or "7 days").	"7 days"
end_date	string	Search mode only. Show posts up to this date (inclusive); relative values count back from today.	"2026-08-25"
max_posts	number	Results per user or search keyword (10-10000). In User Posts, high limits need enough public posts on that account. In Search, this is a cap and available public results may be lower. Find Profiles returns up to 10 account matches per keyword and this setting does not raise that cap.	50

Note: Date options and sort order apply to Search mode only. Leaving one set while you run another mode is fine: it is ignored and the run proceeds normally.

Mode 1: Scrape User Posts (with Profile Data)

Extract posts and available public profile information from a specific user:

{
  "mode": "user",
  "usernames": ["zuck"],
  "max_posts": 50
}

Optional: flag a user's posts by keyword

Add keywords in User Posts, User Reposts or User Replies mode to flag which returned rows are textually relevant to your terms. Nothing is removed; each row gets a keyword_match value you can filter later.

{
  "mode": "user",
  "usernames": ["zuck"],
  "keywords": ["AI"],
  "max_posts": 100
}

Batch Scraping Multiple Users

Scrape posts from multiple users in one run (up to 20 usernames):

{
  "mode": "user",
  "usernames": ["zuck", "instagram", "nike"],
  "max_posts": 50
}

This returns:

• Available public bio and follower count
• Contact information extracted from bio (emails, phones, links)
• Recent public posts with engagement metrics when publicly available (likes, replies, reposts, shares, views)
• Profile verification status
• Media attachments
• Mentions and hashtags: mentions lists usernames found in the text, while mentioned_users and mentioned_accounts list the accounts the platform itself resolved as tagged; the two can differ on a given post

Mode 2: Search Posts by Keyword or Hashtag

Find public posts about a keyword or hashtag. Result volume depends on what Threads shows publicly for that term at run time.

Tip: Include or omit the #: both AI and #AI work. Very specific or cross-language phrases may return fewer public results.

Understanding Search Filters

Filter	Best For	What You Get
Top (top)	Discovering notable content	The posts Threads surfaces as most relevant for that term. Row order is not a popularity ranking.
Recent (recent)	Monitoring new activity	Prioritizes newest matching public posts; sort created_at descending for one combined timeline

Top Results

{
  "mode": "search",
  "keywords": ["artificial intelligence", "#startup"],
  "search_filter": "top",
  "max_posts": 50
}

Recent Posts

{
  "mode": "search",
  "keywords": ["AI"],
  "search_filter": "recent",
  "max_posts": 50
}

Mode 3: Find Profiles (Discover Account Matches by Keyword)

Discover public Threads account matches for a keyword. Each keyword can return up to 10 accounts with public profile fields such as follower count, bio when present, links, and extracted contacts. Some keywords may return fewer than 10. Use related keywords to discover more accounts: an account matched by two of your keywords is still delivered once.

{
  "mode": "profiles",
  "keywords": ["coffee roaster", "specialty coffee"]
}

Output rows are account profiles (not posts), with key fields such as username, display_name, bio, followers_count, is_verified, profile_url, external_links, emails, phones, and search_keyword. Empty bio/contact fields can mean the account did not publish that information.

Every row includes record_type: post for User Posts, Reposts, Replies and Search rows, and profile for Find Profiles rows.

Local search tip: For local account discovery, include the place name in the keyword (e.g. Brooklyn coffee rather than only coffee).

Mode 4: Scrape User Reposts

(beta) Reposts and Replies are newer than the other three modes and carry that label in the input form. They return the same row shape as User Posts plus the extra columns described below, but an occasional run can come back short.

Collect the posts an account boosted, rather than the posts it wrote:

{
  "mode": "user_reposts",
  "usernames": ["natgeo"],
  "max_posts": 50
}

Each row is the original post, with its own author and engagement fields, plus three columns saying who boosted it and when: reposted_by_username, reposted_by_user_id and reposted_at.

If two of the accounts you asked for boosted the same post, you get one row per account, not one row in total.

Mode 5: Scrape User Replies

(beta) The same beta note as Mode 4 applies: same row shape as User Posts plus the extra columns below, and an occasional run can come back short.

Collect the public replies an account wrote:

{
  "mode": "user_replies",
  "usernames": ["mosseri"],
  "max_posts": 50
}

Each row is the account's own reply, with the same fields any other post row carries, plus the post it answered: replied_to_post_code, replied_to_post_url, replied_to_author_username, replied_to_author_user_id, replied_to_text and replied_to_created_at. When the conversation does not name the answered post unambiguously, those fields are left empty rather than guessed.

This mode collects the replies an account wrote, wherever it wrote them. If you want the other direction, the comment thread under one specific post, that is a different tool: Threads Replies Scraper.

Rows from all three user modes carry source_tab (posts, reposts or replies); Search rows leave it empty. So a combined export stays sortable by where each row came from.

Periodic Topic Monitoring

If you want to monitor keywords on a schedule (for lead generation, brand-mention tracking, or competitor monitoring), pair mode: "search" with search_filter: "recent" and run the actor automatically:

{
  "mode": "search",
  "keywords": ["YOUR_KEYWORD_1", "YOUR_KEYWORD_2"],
  "search_filter": "recent",
  "max_posts": 30
}

Suggested workflow

1. Schedule the actor: Use Apify Schedules to trigger regular runs.
2. Review each run's dataset: every post includes a unique identifier and timestamp.
3. Handle your own downstream workflow:
   • Deduplicate against your existing records using each post's unique identifier
   • Optionally filter by post timestamp to keep entries within your time window
   • Trigger your own notifications or follow-up workflow
   • Append new entries to your workspace or spreadsheet

Tips

• Each run returns a finite batch of matching public posts. For high-volume keywords, review the returned timestamps and adjust your schedule to match your monitoring needs.
• The actor does not perform cross-run deduplication, which is intentional, so you keep full control over data retention and storage.

Date Filtering (Search Mode Only)

Narrow search results to a date range. The range can be a recent window (for example the last 7 or 30 days) or an earlier period. Both relative dates ("7 days", counted back from today) and absolute dates (YYYY-MM-DD) are accepted.

{
  "mode": "search",
  "keywords": ["AI"],
  "search_filter": "recent",
  "start_date": "7 days",
  "max_posts": 50
}

Output Format

Viewing Your Data

Use the field-focused dataset views to inspect posts, profiles, engagement fields, or contact fields. Some views are most useful for specific modes. Switch to "All fields" when you need the complete record.

• 🧵 Posts: post content, dates, author, media flags, and search fields
• 📊 Engagement: likes, replies, reposts, shares, views, and analysis flags
• 👥 Profiles: accounts found in Find Profiles mode
• 🔗 Contacts: public profile links and contact fields

Complete Post Record Example

Each post is saved as a separate record. Below is an illustrative example of the fields from user mode. The account, URL, text, dates, and numbers are fictional placeholders.

{
  "record_type": "post",
  "post_url": "https://www.threads.com/@example_creator/post/EXAMPLE01",
  "post_code": "EXAMPLE01",
  "text_content": "Launching a new research note today...",
  "created_at": "2026-01-12T18:00:30+00:00",
  "created_at_display": "2026-01-12 18:00 UTC",
  "created_at_timestamp": 1768240830,
  "like_count": 268,
  "reply_count": 81,
  "repost_count": 12,
  "quote_count": 4,
  "share_count": 22,
  "view_count": 41651,
  "view_count_status": "available",
  "has_media": false,
  "media_type": "text",
  "media_url": "",
  "media_urls": [],
  "media_width": null,
  "media_height": null,
  "has_audio": null,
  "is_ai_generated": false,
  "is_gif": false,
  "is_sticker_post": false,
  "sticker_count": 0,
  "sticker_ids": [],
  "sticker_urls": [],
  "hashtags": [],
  "mentions": [],
  "urls": [],
  "is_pinned": false,
  "is_edited": false,
  "is_paid_partnership": false,
  "accessibility_caption": "",
  "is_reply": false,
  "reply_to_username": "",
  "reply_control": "everyone",
  "is_quote_post": false,
  "quoted_post_url": "",
  "is_repost": false,
  "reposted_post_url": "",
  "repost_code": "",
  "repost_url": "",
  "reposted_by_username": "",
  "reposted_by_user_id": "",
  "reposted_at": "",
  "replied_to_post_code": "",
  "replied_to_post_url": "",
  "replied_to_author_username": "",
  "replied_to_author_user_id": "",
  "replied_to_text": "",
  "replied_to_created_at": "",
  "replied_to_available": null,
  "replied_to_like_count": null,
  "replied_to_reply_count": null,
  "replied_to_repost_count": null,
  "replied_to_quote_count": null,
  "replied_to_share_count": null,
  "conversation_id": "",
  "source_tab": "posts",
  "is_spoiler": false,
  "language": "",
  "topic_tag": "",
  "location_name": "",
  "location_id": "",
  "link_title": "",
  "mentioned_users": ["example_partner"],
  "mentioned_accounts": [{ "username": "example_partner", "user_id": "10000000002" }],
  "scraped_at": "2026-04-25T10:00:00+00:00",
  "username": "example_creator",
  "requested_username": "example_creator",
  "user_id": "10000000001",
  "display_name": "Example Creator",
  "profile_url": "https://www.threads.com/@example_creator",
  "is_verified": true,
  "followers_count": 54000,
  "bio": "Research notes and product updates",
  "profile_pic_url": "https://...",
  "profile_pic_hd_url": "https://...",
  "fediverse_enabled": true,
  "external_links": ["https://example.com"],
  "emails": [],
  "phones": [],
  "bio_links": [],
  "profile_tags": ["AI", "Tech"],
  "podcast_name": "",
  "podcast_url": "",
  "podcast_platform": "",
  "search_filter": null,
  "search_keyword": null,
  "search_keywords": [],
  "keyword_match": null
}

In Search mode, search_keyword remains the first selected keyword for backward compatibility, while search_keywords lists every input keyword that independently selected that deduplicated row. keyword_match is true when the text matches at least one of those selected keywords. A keyword is listed only when it actually selected that row for delivery. In the three user modes, search_keywords is empty; search_filter stays null, and search_keyword plus keyword_match are null unless you supply keywords to tag matching rows.

If a later Search or Find Profiles keyword discovers that an append-only row also matches it, Run Results includes keyword_membership_updates. Match the Dataset row using the entry's match_field and match_value, then add its added_keywords to your analysis. For a username match, compare without case. This preserves every keyword relationship without creating a duplicate Dataset row.

The same rule applies when two User Posts, Reposts or Replies inputs resolve to the same saved result, such as an old handle and its renamed handle. Run Results includes input_membership_updates: match the Dataset row with all fields in match_fields, then add added_usernames to your input-to-result mapping. The Dataset still contains one physical row.

User Posts, Reposts and Replies also include requested_username. It normally matches the requested account; if Threads redirects an old handle after an account rename, requested_username keeps your original input while username continues to identify the actual post author.

What To Expect

Result availability

• Result volume depends on how much public content exists for your query, so different modes and queries return different amounts.
• If a run looks lighter than expected, check Run Results (the run summary on the run's Output tab in the Apify Console), then broaden the query or switch the sort order.
• In a batch run, one username or keyword can fail while others still save results.
• In Search batches, mixed_item_results simply means keywords finished differently in this run. See How to read Search stopping reasons below for how to read it.

Working without a login

This actor reads only what Threads shows publicly, so niche or uncommon keywords can return very few results, or none, even when you can see those posts while logged in. When no_results appears, it means the public search window checked in this run returned no matching rows. It does not prove that no matching post exists anywhere on Threads. If the search was interrupted, the run is reported as partial instead of pretending the empty result was confirmed.

How to read Search stopping reasons

• requested_limit_reached: the run delivered the number you asked for. There may still be more matching posts.
• search_window_limit_reached: the planned search window returned fewer posts than you asked for. Another sort order or a related keyword may find more.
• date_filter_unverifiable: at least one candidate did not include a usable public timestamp, so the Actor could not safely decide whether it belonged inside your date range. The run stays partial instead of claiming the range was empty.
• search_depth_limit_reached: the run reached its collection limit for that keyword before your number. More posts may remain.
• search_time_budget_reached: the time allowed for that keyword ended before reaching your number. More matching posts may remain.
• low_yield_stop: the run stopped because it was no longer finding new posts for that keyword. More posts may still exist.
• mixed_item_results: keywords had different completion outcomes or stopped for different reasons. It does not mean Threads has no more matching posts. Check search_items in Run Results to see which keyword was short and why; this is not proof that the shorter keywords are exhausted.
• budget_reached: the run stopped at the limit you set for it. Everything saved before that point is kept, and a new run with a higher limit can continue past it.
• time_limit_reached / run_stopped_early: the run ended before finishing, for example at its time limit or after being stopped, and wrapped up with what it had.
• temporarily_unavailable: part of the collection could not be completed on this attempt. Running it again can pick up what is missing.

Per-keyword codes in search_items

Each keyword in search_items also reports its own status (complete, partial, or failed) and its own stop_reason. Besides the reasons above, a keyword can report an operational code about this run's attempt: setup_unavailable (search could not be set up for this keyword this time), request_timeout or request_unavailable (a page request ran out of time or could not be completed), save_unavailable or save_incomplete (prepared rows could not be saved, or only partly saved), details_unavailable or details_incomplete (public author details could not be fully attached), or unexpected_error. These codes describe this run's attempt, not the keyword itself; running again usually completes them.

posts_saved counts unique physical Dataset rows. search_items[].results_saved counts confirmed keyword memberships, so one deduplicated post selected by two keywords counts once in posts_saved and once for each keyword. This is expected and does not mean the row was saved twice.

If you need content that is visible only inside your own logged-in account, this is not the right tool.

FAQ

Q: What engagement metrics can I get? A: The scraper extracts likes, replies, reposts, quotes, shares, and views when publicly available. On some posts, share_count and view_count can be null. view_count_status tells you which of three things happened: available (a number was read, including 0), not_public_yet (the post is too recent for a view count to exist yet), or not_checked (the count was not read on this run; run it again to pick it up). In Run Results, view_counts_missing counts the post rows this run checked that had no view count, and view_counts_not_public_yet reports how many of those are simply too recent.

Q: How does contact information extraction work? A: Contact fields depend on the public bio text and profile links. The actor does not read text inside images or on linked pages.

Q: Is there an identifier I can join on besides the username?

A: Yes. Every post author and every profile record carries user_id next to username. A username is the account's handle; user_id is a separate identifier assigned by the platform and is not derived from it. Across our runs the same account returns the same user_id in User Posts, Search and Find Profiles alike, so it is the field to join on when a handle alone would be ambiguous. When it could not be resolved for a record it is null, never an empty string, so it is always clear which case you are looking at.

Q: How do I see the complete field set? A: In the Apify console, switch to "All fields". You will see the complete record fields, including fields that may be null for some records.

Q: Does it support batch scraping of multiple users or keywords? A: Yes. You can enter up to 20 usernames or 20 keywords in one run. Results are combined into the same dataset.

Q: Can I tell which posts are sponsored / paid partnerships? A: Every post includes an is_paid_partnership field. true means a paid-partnership or branded-content signal is publicly shown for that row. false means that signal was not present in the public data returned for that row.

Q: Can I see quote posts and reposts, and the original they reference? A: Yes. Quotes are marked on the row itself with is_quote_post and quoted_post_url. Reposts come from 🔁 User Reposts mode, where each row is the original post and reposted_by_username, reposted_by_user_id, reposted_at and source_tab say who boosted it and when.

Q: Can I distinguish top-level posts from replies? A: Yes, when reply context is present in the returned public data. is_reply marks replies, reply_to_username shows who is being replied to, and reply_control reflects the reply-audience setting when available.

Q: What about image alt text and media size? A: When available, accessibility_caption contains an image description; otherwise it is an empty string (""). media_width and media_height contain the available media dimensions in pixels and are null when not available (for example text-only posts).

Q: Can I tell whether a post has a public AI-content label or whether a video has sound? A: is_ai_generated is true when the post carries an AI-content label, false when the checked public post has no such label, and null when it cannot be determined. has_audio is true/false for checked video posts and null for non-video posts or unavailable detail.

Q: Is there a Fediverse / ActivityPub signal? A: Yes. fediverse_enabled is true when a Fediverse sharing indicator is publicly shown on the profile; otherwise it is false or null, depending on whether profile details were available for that row.

Q: Do I need to choose between hashtag and full-text search? A: No. Enter your keyword with or without #, then pick a top or recent sort order. Results depend on what Threads makes publicly available.

Q: How does Find Profiles work, and how many accounts can it return? A: Find Profiles (mode: "profiles") returns up to 10 public account matches per keyword with profile fields such as follower count, bio when present, links, and contacts. Some keywords may return fewer than 10. To reach more accounts, use several related keywords.

Q: How can I find local accounts? A: Put the place name directly in the keyword (e.g. austin coffee) so the search phrase is what the matching is based on. Geo-targeting is not supported.

Q: Why do search results seem unrelated to the keyword? A: Threads search ranking is not strict keyword-only matching. Results can include broadly related content and sometimes unrelated content. Use keyword_match as a text-relevance aid in downstream filtering; it is true when the row text matches at least one keyword in search_keywords, but it is not an exact-match guarantee.

Q: Why do I get fewer results than my max_posts limit? A: max_posts is a ceiling, not a promise. One run checks a bounded public search window, so it can return fewer rows even for an active keyword. Read collection_stop_reason in Run Results: it names why the run stopped (the full list is under How to read Search stopping reasons above). None of the short-result reasons means every matching post on Threads was checked. Threads also limits how much of a search a visitor who is not logged in can see, and that visible window changes over time; when it is what ended the run, the run is marked partial rather than complete. You can try a broader keyword, another sort order, a wider date range, or split the search into related keywords.

Q: Why are bio or followers_count sometimes empty in Search mode? A: Some accounts expose limited public profile fields in search results. When a field is not publicly available for that account at run time, it may be empty.

Explore Threads Workflows

• Threads Replies Scraper: collect public replies and comments for conversation analysis.
• Threads 海巡 - 貼文、個人檔案與關鍵字搜尋: the Traditional Chinese edition of this Actor, with a zh-TW interface and output labels.

Disclaimer

This Actor is provided for research and educational purposes. You are solely responsible for how you use it and for ensuring that your use of the Actor and any data it returns complies with all laws and regulations applicable to you. Data availability, completeness, and accuracy may vary over time and between runs. The developers accept no liability for how this tool or the data it returns is used.

---

Keywords: Threads scraper, Threads search scraper, Threads data scraper, Meta Threads scraper, Threads post extractor, Threads profile scraper, Threads hashtag search, Threads keyword search, Threads search without login, Threads data extraction, export Threads posts, download Threads data, Threads exporter, Threads posts, Threads engagement metrics, follower count scraper, Threads bio email extractor, Threads account finder, influencer discovery, Threads influencer research, social media scraper, social media monitoring, brand monitoring, competitor analysis, competitor intelligence, trend research, content research tool, social listening tool, market research, Instagram Threads, public Threads data, Threads automation, Threads search, OpenClaw, Claude Code, Gemini CLI, Antigravity, Codex, ChatGPT