Threads Replies Scraper - No Login

Scrape every reply Threads shows publicly on a post - reliably and fully populated. No account, no cookies, no API key. Paste post URLs, get back structured JSON: author, text, engagement, media, nesting, the lot.

Built because the only other replies scraper on the Store finishes 4 out of 5 runs and ships thin, partly-empty records. This one is built reliability-first: automatic recovery from soft-blocks, health-monitored around the clock, and it never silently serves garbage - a bad run pauses itself rather than damage your data.

Features

• No login required - works on any public Threads post, exactly as a logged-out visitor sees it
• Full field set - author (username, display name, verified badge, avatar), engagement (likes, replies, reposts, quotes, shares), media (photo/video/carousel/gif, dimensions, alt text), timestamps, mentions, hashtags, links
• Real reply nesting - is_top_level and parent_reply_id resolved for ~96% of replies; the ambiguous ~4% (usually an author replying to themselves) are left null, never guessed
• Honest nulls - a field is null only when Threads genuinely doesn't expose it logged-out (e.g. view_count), never because parsing failed
• Partial-batch delivery - one broken post in a 20-post batch never blocks the rest; you get every post that worked, the failures are flagged, and you're billed only for what's delivered
• Self-healing - automatic session rotation if Threads serves a soft-block, a rotating canary that watches run health around the clock, and an auto-pause if something breaks for real (protects you from a degraded run, not just us)

Input

Field	Type	Default	Description
post_urls	array of strings	(required)	Up to 20 Threads post URLs, e.g. https://www.threads.com/@username/post/ABC123
max_replies	integer	40	Cap on replies returned per post. Range 10-50, matching Threads' own logged-out visible ceiling
include_nested_replies	boolean	true	Adds is_top_level and parent_reply_id to each reply

Example input:

{
  "post_urls": ["https://www.threads.com/@zuck/post/C8Vk2tLxupR"],
  "max_replies": 40,
  "include_nested_replies": true
}

Example output

One reply record, taken from a live run (image URL truncated for readability):

{
  "reply_id": "Dad19aBFN5B",
  "author_username": "brycebots",
  "author_display_name": "Bryce Botsford",
  "author_is_verified": false,
  "author_profile_pic_url": "https://instagram.frba3-2.fna.fbcdn.net/...",
  "author_profile_url": "https://www.threads.com/@brycebots",
  "text_content": "They literally made fun of AI in the movie",
  "mentions": [],
  "hashtags": [],
  "urls": [],
  "like_count": 149,
  "reply_count": 0,
  "repost_count": 0,
  "quote_count": 0,
  "share_count": 1,
  "view_count": null,
  "created_at": "2026-07-06T21:20:25.000Z",
  "has_media": true,
  "media_type": 1,
  "media_category": "photo",
  "media_url": "https://instagram.frba3-2.fna.fbcdn.net/...",
  "media_width": 1290,
  "media_height": 1573,
  "reply_to_username": "lanenomad",
  "parent_reply_id": null,
  "is_top_level": true,
  "is_reply": true,
  "item_type": "reply",
  "source_post_url": "https://www.threads.com/@lanenomad/post/DadD0tukYzZ",
  "scraped_at": "2026-07-13T05:38:33.178Z"
}

The original post is returned too (item_type: "original_post"), so you always have the full context, not just an orphaned list of replies.

How it works

We fetch the exact page a logged-out browser loads when you open a public Threads post. Threads embeds the entire reply tree directly in that page's HTML - no separate API call, no login. We parse it straight out of the response.

If Threads ever serves an empty app shell instead of the real page (a rare, intermittent soft-block), we detect it immediately and retry through a fresh residential IP - the kind of self-healing most scrapers in this niche don't bother with.

Pricing

$5 per 1,000 results, and nothing else - no run fee, no separate platform usage charge. A result is either a reply or the original post - so a 20-post batch averaging ~40 replies each returns roughly 800 results, about $4.00. On Apify's free plan your monthly platform credit covers your first ~1,000 results, so you can check the data quality before paying anything.

How we compare to the incumbent

	This actor	The only other replies scraper on the Store
Rating	new	1.19★
Success rate	reliability-first design: session rotation, canary monitoring, auto-pause	77.9% and degrading
Max replies/post	10-50	10-50
Nesting	is_top_level + parent_reply_id, ~96% resolved	flat, reply_to_username only
Field completeness	full engagement + media + honest null discipline	partial
Price	$5 / 1,000 results	$2.50 / 1,000 results

We cost more per result because we finish the run and give you real data when we do. A cheap actor that fails one run in five, or hands back half-empty records, isn't actually cheaper.

Limitations (read this before you buy)

Threads caps how many replies it shows a logged-out visitor - typically tens per post, up to this actor's 50-per-post ceiling, never the full count on a large thread. This actor returns every reply in that public view, completely and reliably. It does not, and cannot without logging in, return every reply on a thread with hundreds or thousands of replies. Any actor claiming otherwise is either exaggerating or scraping while logged in, which is a different risk profile (account bans, ToS exposure) than this actor takes on.

A small share of replies (~4%, usually a self-reply) can't be confidently linked to a parent reply. We leave parent_reply_id as null in that case rather than guess - a wrong guess is worse than an honest gap.

FAQ

Does this work without a Threads account? Yes. Every request is made logged-out, exactly as any visitor sees it. No cookies, no session, no login step.

Why didn't I get every reply on my post? See Limitations above - Threads itself caps what a logged-out visitor can see. We return everything in that view.

What counts as one billable result? Each reply and the original post, each as one dataset record.

Can this scrape a private or protected account's posts? No. It only reads what's publicly visible, by design - that's also why it never needs your Threads credentials.

What happens if one of my post URLs is deleted or fails? The batch still succeeds. The working posts are delivered, the failed one is flagged by URL in the run summary, and you're only billed for what came back.

How long does a run take? Roughly 5-10 seconds per post, run sequentially (deliberately, not in parallel - it's the safer way to stay reliable at volume).

Use in Claude, ChatGPT, or any MCP-compatible agent

Apify actors are callable through Apify's MCP server, so this scraper can be wired straight into an AI agent's toolset - hand it a Threads post URL in conversation and get structured reply data back, no separate scraping step.

Changelog

• v0.1 - initial release: fetch, parse, reliability layer, honest field completeness vs the incumbent.

---

This actor is not affiliated with, endorsed by, or sponsored by Meta Platforms, Inc. or Threads. It reads only publicly available data through Threads' own public web pages, exactly as a logged-out visitor's browser would.