Boosty Scraper — Creators, Subscribers & Tier Prices

Boosty is where a large part of the Russian-speaking creator economy actually earns. This Actor reads any public Boosty creator: how many followers they have, what every subscription tier costs, what their funding goals are, how often they post, how much engagement each post gets, and how much of their output sits behind the paywall — then tells you when any of that moves.

Public API, no key, no login.

What you get per creator

Audience	Follower count, total posts, showcase items
Pricing	Every tier with its price in the creator's own currency and USD/EUR/RUB, what the tier promises, which tiers unlock the Discord role or private Telegram, plus hidden and limited flags
Live discounts	Which tiers are running a promo right now, how deep, until when — and how many people have taken it
Goals	Open funding goals with current, target and progress percent
Output	Posts in the last 30/90 days, days since the last post, dormant flag
Engagement	Average reactions, comments and total engagement per post, and every reaction type broken out
Paywall mix	How many posts are free versus gated, and which tier gates each one
Comments	On free posts: author, text, date, replies and reactions, plus who comments most
Content	On free posts: the full post text, plus images, videos, audio and outbound links
Reach	Linked Instagram, Telegram, YouTube, VK, Twitch, TikTok, X, plus the Discord server behind the paid tiers

For a paid post only metadata is read: its body is not served to anonymous callers and is never requested, so you get the title, the date, the engagement and the tier that gates it. For a free post you also get the text and the media the creator published.

The one hard number about money on this platform

Every subscription discount carries count.activation — how many people actually took it. One creator in the sample runs four promos at 10–25% off with 71 activations between them. That is not an estimate and not a follower count: those are people who paid at a named price.

Treat it as a floor, and only that. Activations accumulate over the life of the promo, so some of those people may since have cancelled, and subscribers who joined at full price are not counted at all. The Actor reports the activations and the discounted price and stops there; multiplying them is your call, not a number it will invent for you.

Two things this refuses to guess

No revenue estimate. Boosty publishes followers, not paying subscribers. Multiplying followers by a tier price would produce a confident-looking number that is wrong by an unknown factor. You get the real inputs — follower count, every tier price, goal progress — and you decide.

"Likes" is not what it says. Boosty's likes counter is, in fact, the ❤️ reaction: it matched reactions.heart in all 180 posts sampled across six blogs, while the real 👍 lives in reactions.like and is usually the larger number. Reporting it as "likes" would understate a creator badly — one blog in the sample averages 0.77 hearts and 14.8 total engagement. So the field is called hearts, and avgEngagement (all reactions + comments) is the headline number.

Comments, and what the paywall hides

Comment threads are public on free posts only — the thread of a paid post answers HTTP 403 to anonymous callers, even though that post's own title, date and engagement are public. So comments are requested only for free posts that actually have any, which also means a creator whose output is entirely paid costs no extra requests at all. commentsLockedPosts says how many threads were behind the paywall, so an empty comment list is never mistaken for a silent audience.

Media links, and which of them actually work

Handing over a link that 403s is worse than handing over none, so all three kinds were tested rather than assumed:

	Works for you?
Images (images.boosty.to)	Yes — unsigned and durable, opens for anyone
Video watch page (ok.ru/video/<id>) and its preview image	Yes — durable, opens for anyone
Video stream (streamUrl)	Only from the machine that scraped it. OK.ru signs these with an expires timestamp and the requesting IP, so it works during the run and not afterwards from your desk. Flagged streamUrlIsSigned.
Audio (cdn.boosty.to)	No — 403 to anonymous callers with or without Referer/Origin. The URL is an identifier, not a download, and is flagged urlNeedsSession. Title, artist, duration, size and format are all still there.

Monitor mode

Run it on a schedule against the creators you follow, compete with, or sponsor.

changeType	Means
tier_change	A tier's price moved, or a tier was added or removed — tierPriceChanges spells out each one, in the creator's own currency
promo_change	A discount started or ended — the list price does not move, so this would otherwise pass as unchanged
new_posts	They published since the last run — newPosts lists which ones, with title, link, date and the tier that gates each
went_quiet	The run where they crossed your dormancy threshold — announced once, not repeated every run
subscribers_up / subscribers_down	Audience moved, with subscribersDelta
goal_progress	A funding goal moved or changed
no_longer_available	Boosty no longer serves the blog — deleted, renamed or blocked
new, unchanged	First sighting / nothing moved

A creator that 404s but was never in your baseline is reported as a probable typo, not as news. The disappearance alert only fires for a creator that was really there before.

Two different numbers describe new output, deliberately. newPostsSinceLastRun is the change in the creator's own post counter — it can read zero when they delete one post and publish another. newPosts is the list actually observed, matched by publish date against the last run, and is the one to trust. If every post read was newer than the baseline the creator may have published more than your post limit could see, and newPostsTruncated says so instead of implying the list is complete.

Input

Field	What it does
Creators	Profile URLs, post URLs or bare handles — https://boosty.to/welovegames, welovegames, @welovegames all work and de-duplicate.
Include tiers / goals	One extra request each per creator. Tiers bring the live discounts with them.
Include comments / Comments per post	Reads threads on free posts. Nothing is requested for posts with no comments or behind a paywall.
Analyse posts / Posts per creator	How far back to read. More posts means truer averages.
Include the individual posts	Adds every post read: title, date, reactions, comments, gating tier, media types, tags.
Treat as dormant after	Your definition of quiet, not a hidden constant.
Monitor mode / Output only changes	The alert behaviour above.

Notes & limits

Properties of the source, measured rather than assumed:

• A tier's bare price field is EUR, whatever the creator actually charges. One tier reads price: 1.42 while the creator's own price is 130 ₽. Every price here is reported with its currency, so nothing is quoted in a unit nobody named.
• The USD and EUR figures are converted live and drift. Two calls two seconds apart returned 1.65 then 1.62 USD for a tier whose rouble price never moved. They are fine for comparing creators; they are useless for detecting change, so monitor mode diffs the creator's own currency and never the converted number. Otherwise almost every run would report a price change that never happened.
• Post paging uses a cursor, not a row number. An integer offset answers HTTP 400 on every page, which looks like a dead endpoint. The end of a blog is an empty page, never a short one.
• Boosty's search and feed endpoints require a login (HTTP 401), so this Actor works from creators you name rather than discovering them. Handles come from the creator's own links on YouTube, Telegram, Twitch and so on.
• The 30- and 90-day post counts are left empty when the posts read do not span that window — reporting "3 posts in 90 days" because only 20 posts were fetched would be an artefact of your own limit, not a fact about the creator. Raise the post limit to fill them in.
• Comments on paid posts are not public (HTTP 403) while the post metadata is. Requesting them anyway would burn one request per paywalled post and make every creator look "partial".
• Comment reactions ship both laugh and laught — the same reaction under two spellings. Summing the list blindly double-counts it.
• Discovery is genuinely closed, checked three ways: the search and feed endpoints answer 401, oauth/token/ only issues tokens against a real account's refresh token, the search page is a client-side app with nothing in its HTML, and robots.txt disallows *?searchQuery=*. This Actor therefore works from creators you name.
• A blog that requires a login to view is reported as such, not silently dropped.
• Load-tested at real scale: 75 creators in 24 s at 4 parallel requests, 13 s at 8, 28 s with comments, 40 s reading 150 posts each — no throttling and no warnings from Boosty.
• The Actor honours your Maximum cost per run and stops when the cap is reached.