Pinterest Rich Pin, Board & Creator Field Scraper

Pinterest Rich Pin, Board & Creator Field Scraper

Pinterest's own pages carry far more structured data than a basic Pin scrape surfaces — media variants, full creator and board objects, product and SEO metadata, Story Pin pages, and reaction breakdowns. This actor pulls that deeper field set for Pins, profiles, boards, search results, and public comments, without requiring login cookies; anything Pinterest doesn't publish for a given record is simply left out of the newer camelCase fields.

What this actor does

• Ten collection modes plus auto-detect: Pin detail, Pin comments, Pin search, video search, profile detail, a creator's Pins, a creator's boards, board detail, and board Pins, plus auto for pasted links.
• Rich Pin context: media variants, creator and original-creator objects, board objects, destination links, product and SEO metadata, Story Pin pages and reaction counts when Pinterest supplies them.
• Pin filters: saves, creator followers, verified creator, destination link, image-only, promoted Pins and shoppable Pins. They apply to Pin rows only, never to profile, board or comment rows.
• Cursor paging: follows Pinterest's own next-page cursors for searches, creator feeds and boards, and skips Pins already returned earlier in the run.
• Encrypted feed IDs: search and board feeds can return an opaque e1_... Pin ID instead of a numeric one. Those rows are kept as list records with idIsEncrypted: true, detailAvailable: false and no Pin URL, so a run is never silently empty.
• Numeric-ID recovery: for a creator feed that returns encrypted IDs, the actor reads the creator's public RSS feed (or the numeric user-pins feed) and emits openable numeric Pin rows when resolveNumericIds is on.
• Run provenance: every row carries sourceType, sourceValue, position, seed, and query on search rows, plus numericIdSource when a numeric ID was recovered.
• Run summary: the default key-value store gets a RUN_SUMMARY with raw rows, accepted rows, rejected rows and the reason a run returned nothing.
• Omitted fields: the camelCase fields are dropped when Pinterest doesn't supply them. A few older snake_case fields fall back to 0, false or "public" instead (see the note under the output section).

Output per record

Every row includes record_type/recordType (pin, profile, board or comment), source_url (the search, profile, board or Pin page it came from), scraped_at, sourceType, sourceValue and position. Search rows also carry query; every row carries seed. Rows produced from a recovered numeric ID carry numericIdSource (userPins or profileRss).

Pin fields

Many Pin values appear twice: once under an older snake_case name and once under a newer camelCase name. Any field below may be absent when Pinterest doesn't expose it for that Pin.

• Identity & text: id/pinId, idIsEncrypted, detailAvailable, commentsAvailable, url/pinUrl (numeric Pins only), title, description, altText, created_at/createdAt, dominant_color, link/linkUrl, linkDomain, domain, contentCategory, canonicalUrl, canonicalPinId, canonicalPinUrl
• Media: images (size → URL map), imageUrl, imageVariants, videoVariants, video_url/videoUrl, is_video/isVideo, mediaType (image, animatedImage or video), pinFormat (story for Story Pins), mediaMethod, storyPinId, storyPages
• SEO & link tracking: seoTitle, seoDescription, seoUrl, mobileLinkUrl, trackedLinkUrl, utmLinkUrl
• Creator: creator, creatorId, creatorUsername, creatorName, creatorFollowerCount, originalCreator (only when it differs from creator), pinner_id, pinner_username, pinner_name, pinner_followers
• Board: board, boardId/board_id, boardName/board_name, boardDescription, boardUrl/board_url, boardCoverImageUrl, boardPinCount, boardFollowerCount, boardSectionCount, boardCollaboratorCount, boardPrivacy, boardCreatedAt, boardIsCollaborative, boardFollowedByMe
• Engagement: saves/save_count, commentCount/comment_count, repinCount, shareCount, likeCount/like_count, reactionCount/reaction_count, reactionCounts/reaction_counts
• Status flags: isPromoted, commentsDisabled, commentsEligible, isEligibleForImageDownload, isEligibleForProduct, isOutOfStock, isStaleProduct, isUnsafeForComments, isDownstreamPromotion, shouldMute
• Shopping: product, productPrice, productCurrency, productMerchant, productAvailability, isShoppable, shoppingFlags, richMetadata

creator and originalCreator can carry id, username, fullName, firstName, followerCount, avatarSmallUrl, avatarMediumUrl, avatarLargeUrl, isVerified, isAdsOnly, isPrivate and followedByMe. Each imageVariants/videoVariants entry is keyed by Pinterest's own size name and can carry url, width, height, duration and thumbnailUrl. Story Pin pages can carry pageId, layout and a block list with type, text, images and videos.

Profile fields

• Identity: type (always user), id/profileId, username, full_name/fullName, bio, location, createdAt
• Counts: follower_count/followerCount, following_count/followingCount, pin_count/pinCount, board_count/boardCount, monthlyViews
• Media & links: avatarUrls, avatarUrl, image_url, website_url/websiteUrl, domainUrl, profile_url/profileUrl
• Account flags: is_verified, isVerified, isPrivate, isPartner, isVerifiedMerchant, isDomainVerified, isTastemaker

Board fields

• Identity: type and recordType (both board), id/boardId, name, description, url, owner_username/ownerUsername, owner_name, createdAt
• Counts: pin_count/pinCount, follower_count/followerCount, sectionCount, collaboratorCount
• Media & flags: coverImageUrl, privacy, isCollaborative, followedByMe

Comment fields

• Content & identity: id, text, created_at, type, node_id, tags, tagged_users, media, pin_id, pin_url
• Author: author_name, author_username, author_id, author_first_name, author_image, author_image_medium, author_node_id, author_is_private_profile
• Engagement & state: like_count, reaction_counts, reaction_by_me, helpful_count, marked_helpful_by_me, reply_count, is_edited, is_highlighted_by_pin_owner, is_translatable

Note on absent values: every field is written only when Pinterest supplied it for that record. An explicit 0 or false from Pinterest is kept; an unknown engagement count, follower count or video flag is omitted instead of being replaced with a made-up zero. idIsEncrypted, detailAvailable and commentsAvailable are always present because the actor can determine them from the ID shape.

Encrypted Pin IDs

Pinterest increasingly returns an opaque ID such as e1_AfqPShx7dj... from search and board feeds. That ID is real, but pinterest.com/pin/<id>/, the Pin detail endpoint and oEmbed do not open it. The actor keeps the row:

• id and pinId contain the opaque ID exactly as Pinterest returned it.
• idIsEncrypted is true and detailAvailable is false.
• aggregatePinId is emitted when Pinterest supplies the aggregate comments ID. In that case commentsAvailable is true and includeComments can fetch the public comment thread without a Pin URL.
• url and pinUrl are omitted because no public Pin page exists.
• All list fields Pinterest did return remain: title, description, images, creator, board, reaction counts, saves (when the feed carries them), created_at, domain, link, video/story data and so on.

The actor never invents detail-only fields for these rows. If the same Pin is later found through a creator feed with a numeric ID, that row is openable and can be enriched.

Numeric-ID recovery

UserPinsResource normally returns numeric IDs. When it returns only opaque IDs or no rows, resolveNumericIds (default on) makes the actor read the creator's public /<username>/feed.rss. RSS items carry a numeric Pin link, title, description, image and publication date; they are emitted with sourceType: "profileRss" and numericIdSource: "profileRss". RSS contains only the creator's most recent Pins (about 21-25). The user-pins cursor is always preferred when it returns numeric rows.

Commerce fields

When Pinterest supplies a real product payload (product_pin_data, product_metadata or rich_summary.products[].offers), the actor normalizes:

• productPrice in major currency units (Pinterest's price_value is in cents: 1364 -> 13.64).
• productCurrency (upper-case ISO code), productMerchant, productAvailability (in_stock, out_of_stock, preorder, backorder or stale).
• isShoppable only when Pinterest published real product/shopping evidence. A {type, id} metadata stub alone is not treated as shoppable.

The original product object is kept for traceability.

Run summary and zero-row behavior

Every run writes RUN_SUMMARY to the default key-value store:

{
  "rawRows": 20,
  "accepted": 5,
  "rejected": 0,
  "recordsByType": { "pin": 5 },
  "sources": [
    { "sourceType": "search", "sourceValue": "cats", "rawRows": 20, "accepted": 5, "rejected": 0, "rejectReasons": {} }
  ]
}

When a valid source returns rows but none can be emitted, the run fails with a reason such as opaque_id_not_openable or source_returned_no_rows. When filters deliberately remove every row, the run warns with all_rows_filtered_out. The status message always shows raw / accepted / rejected / reason, so a zero-row run is never reported as a plain success.

Input

At least one source is required: keywords, startUrls, pinUrls, usernames, boardUrls, or an explicit mode with its matching source. The auto prefill is search=cats; Apify's validator cannot express "one of", so the actor documents the fallback.

Example: detailed Pin

{
  "mode": "pinDetail",
  "pinUrls": ["https://www.pinterest.com/pin/142567144448888376/"]
}

Example: creator feed with refreshed counts

{
  "mode": "userPins",
  "usernames": ["crazylaurablog"],
  "maxItems": 2000,
  "endPage": 0,
  "enrichDetails": true
}

Example: board Pins with comments

{
  "mode": "boardPins",
  "boardUrls": ["https://www.pinterest.com/nasa/hubble-images/"],
  "maxItems": 500,
  "includeComments": true,
  "maxComments": 20
}

Example: filtered video search

{
  "mode": "searchVideos",
  "keywords": ["easy dinner", "travel guide"],
  "excludeAds": true,
  "maxItems": 500
}

Use cases

• Audit the destination links, SEO titles and alt text on a brand's Pins.
• Compare creators by follower counts, monthly views and verification flags.
• Inventory a board's Pins together with their section and collaborator counts.
• Collect Story Pins and video Pins with every published media size for a creative library.
• Read the comment threads under high-save Pins in a niche.
• Track which Pins in a search are shoppable, out of stock or promoted.

FAQ

Do I need to log in?

No. Every mode works signed-out. The cookie inputs are optional and no mode depends on them.

What happens if I run it with no input?

When no search phrase, URL, username or board is given, the actor collects the single example Pin from the input form so the run still produces a record.

Why were my usernames or boards ignored?

In auto mode the actor only reads keywords/search, startUrls and (when startUrls is empty) pinUrls. Pick userProfile, userPins, userBoards, boardDetail or boardPins to use usernames or boardUrls, or paste the profile and board links into startUrls.

What is the difference between searchPins and searchVideos?

Both run the same Pinterest Pin search. searchVideos then keeps only the results Pinterest marks as video, so it can page through more results to fill maxItems. search is an older name for searchPins.

Can a run return more than 10,000 rows?

Yes. maxItems goes up to 20,000 (or 0 for no limit), and the actor keeps following Pinterest's cursor as long as Pinterest returns more Pins.

Why can a feed contain fewer rows than the account's visible Pin count?

Signed-out feeds, cursor limits, privacy settings, duplicate Pins and your filters all shrink the result. For boards, the actor loads up to maxItems Pins first and then applies the filters, so a filtered board run can return fewer rows than maxItems.

Why do some rows have no url or detail counts?

Search and board feeds can return an encrypted e1_... Pin ID. Pinterest does not open that ID as a Pin page and does not return its detail. Those rows are still real list records: idIsEncrypted is true, detailAvailable is false, and url/pinUrl are left out. The list fields Pinterest did publish remain. When Pinterest supplies an aggregate comments ID (aggregatePinId), the actor can still fetch the public top-level comments for that row. Set includeCommentReplies to fetch depth-one replies for comments that report replies; each reply carries parentCommentId and depth: 1. Deeper nesting is not exposed by the public resource.

What does resolveNumericIds do?

For a creator feed, it enables the numeric-ID recovery path. If the creator's Pin list is empty or contains only encrypted IDs, the actor reads the creator's public RSS feed and emits numeric Pin rows there. RSS is recent-only; set the option to false if you want only the IDs the creator feed itself returns.

Why is a run marked failed with zero rows?

The actor now treats a zero-row run as a bug when the source returned rows it could not emit. The status message and RUN_SUMMARY show rawRows, accepted, rejected and a reason such as opaque_id_not_openable or source_returned_no_rows. A run that is empty because your filters removed every row warns with all_rows_filtered_out instead of failing.

Why is productPrice missing on most Pins?

Pinterest publishes product payloads only for a small subset of Pins. The actor fills productPrice, productCurrency, productMerchant, productAvailability and isShoppable only when Pinterest supplies real product data or shopping flags. A metadata stub with just a type and ID is not treated as a product.

What does enrichDetails change?

For Pins found through search, creator or board feeds, it opens each Pin's own page and overwrites the comment, reaction and save counts, the creator id, name, username and follower count, and the video flag and URL with the Pin page's values. It does not add new media, product or SEO fields, and it slows the run down. pinDetail, comments and Pin URLs in auto already read the Pin's own page.

maxItems counts Pins/source records. Comment rows, including opt-in replies, are additional and are capped by maxComments and maxCommentReplies; a comments run can therefore produce more dataset rows than maxItems.

Are filters based on guessed engagement?

No. minSaves, verifiedOnly, commerceOnly and the other filters only pass a Pin when Pinterest supplied the value they test. A Pin with no published save count is dropped by minSaves.

Which Pinterest links are recognised?

Pin, profile, board, board-section, Ideas, /search/pins/?q=, /search/users/?q= and /search/boards/?q= links on pinterest.com and its regional domains (such as pinterest.co.uk, .de, .fr, .ca, .com.au and .jp). Short pin.it links are not recognised.

What do the newer modes return?

• searchUsers and searchBoards return public people/board search results with cursor pagination.
• ideas returns the Pins in an Ideas feed (list-only; the public feed has no detail/comments).
• A three-segment board-section URL reads BoardSectionsResource + BoardSectionPinsResource, which is the best cookieless source of numeric Pin IDs.
• Profile tabs are routed honestly: pins/_created read created Pins, boards reads the person's boards, and unsupported tabs (_saved, followers, following, likes, activity) warn instead of returning unrelated Pins.
• Related-Pins mode is not exposed because the public resource worked locally but returned no rows from the cloud exit. The e1_... to numeric mapping remains a hard limit; use board sections, user Pins and creator RSS for numeric recovery.

Why do some outbound linkUrl values return 403?

The outbound link is exactly what Pinterest supplied. Etsy short links (etsy.me) can reject clean non-browser clients with 403; the destination is still valid in a browser. Pin media, profile and board URLs were verified HTTP 200.

Is this affiliated with Pinterest?

No. It's an independent tool that reads publicly visible Pinterest data.