Pinterest Scraper — Pins, Boards, Search, Video URLs | $1.99/1k

Turn any Pinterest search into a spreadsheet of pins you can actually work with: title, description, the outbound destination link, a direct image URL with its real pixel dimensions, the dominant colour, the board it lives on, when it was created and whether it is a video. Built for e-commerce and trend researchers, content and social planners, designers hunting reference, and anyone feeding a creative pipeline or an AI agent.

Scrape Pinterest pins by keyword - one keyword or a whole list in a single run - or switch scope to boards and get the boards that rank for a topic instead. Paste a board URL to get every pin saved on that board, or a pin URL to get one detailed row with save, repin, share and reaction counts and the direct video URL of a video pin. No login, no API key, no cookie and no browser: the Actor reads the same public responses pinterest.com loads for a visitor, so a 20-result preview finishes in about 7 seconds and 1,000 pins cost only a handful of HTTP requests. Results export to CSV, JSON, Excel or XML.

This is an unofficial community Actor. It is not affiliated with, endorsed by or connected to Pinterest.

What does this Pinterest scraper do?

You give it keywords or Pinterest URLs; it gives you back one clean row per pin or board.

• Keyword search. One keyword in query, several in queries, or Pinterest search URLs. Scope pins (the default) returns individual pins with their image, description, destination link and board. Scope boards returns the boards that match the keyword, with the board name, description, cover image and pin count.
• Board URLs. https://www.pinterest.com/<owner>/<board>/ returns the pins saved on that board, page after page, up to your limit.
• Pin URLs. https://www.pinterest.com/pin/<id>/ returns one detailed row: everything a search row has, plus save, repin, comment, share and reaction counts and, for a video pin, the MP4 URL.
• Video, domain and alt text on every path. videoUrl for video pins, domain (the host of the outbound link) and altText are filled wherever Pinterest publishes them.

Country domains work (pinterest.co.uk, pinterest.de, ...), and a bare keyword typed into the URL field is accepted as a shortcut. A board or pin that does not exist becomes one free not_found row. Profile URLs and /search/users/ are refused on purpose, each with one free explanatory row, because this Actor does not collect people.

Result status (tri-state output)

Every row says what happened to it, so nothing ever fails silently:

status	Meaning	Billed?
ok	Data was extracted.	Yes
not_found	Pinterest answered normally, but there is no such pin or board, or the search matched nothing.	No
error	The source could not be read or the input entry was unusable; error explains why.	No

You are only charged for ok rows. A run that finds nothing costs you the Actor start event and nothing else.

Use cases for Pinterest search data

• E-commerce trend research. A merchandiser pulls 500 pins for "quiet luxury kitchen", groups them by dominantColor and createdAt, and sees which palettes and product shapes are surfacing this quarter before committing to a buy.
• Content and campaign planning. A social media manager searches their category weekly, reads the title and description wording that keeps appearing, and writes pin copy and hashtags that match how people actually search.
• Creative inspiration and moodboards. A designer or art director collects imageUrl plus imageWidth and imageHeight for a theme and drops the set straight into a moodboard tool, a brief, or a vision-model pipeline — no manual right-click-save.
• Traffic and affiliate research. An affiliate or SEO analyst filters to rows where link is present to see exactly which shops, blogs and recipe sites Pinterest is sending category traffic to.
• Competitor category monitoring. A brand team searches a competitor's product terms and watches which boards their pins get collected on and how large (pinCount) those boards are.
• AI agents and automation. Rows are flat JSON with an explicit status column, so an agent building a mood report or a product feed can tell "no results" apart from "could not check" without parsing prose.

What data you get from every pin or board

One row per result: a pin found by a search, a pin saved on a board, or a pin given by URL. resultType is pin normally, or board when you search with Scope set to boards. Missing means Pinterest did not publish it — fields are omitted, never filled with a guess.

The result

Field	Type	Description
resultType	string	pin for a pin result, board for a board result.
pinId	string	Pinterest's numeric id for the result: the pin id, or the board id when Scope is boards.
title	string	Pin title. When the pinner left it blank, Pinterest's grid title or its image alt text is used instead. For a board row, the board name.
description	string	Description written by the pinner, or the board description. Frequently empty on Pinterest.
pinUrl	string	Canonical pinterest.com URL of the pin, or of the board for a board row.
link	string	Outbound destination URL the pin points at — the shop, recipe or article. Absent for pins uploaded straight to Pinterest.
domain	string	Host of link without a leading www., e.g. amazon.com. Absent when there is no outbound link.
isVideo	boolean	True when the pin plays a video: a video pin, or an idea pin with a video block.
videoUrl	string	Best playable URL of a video pin: the largest MP4 when Pinterest publishes one (a pin URL does), otherwise the HLS playlist (.m3u8). Absent for image pins. The video file itself is never copied.
createdAt	string	ISO 8601 timestamp of when the pin was created. Pinterest does not publish this for boards, nor for the pins in a board listing.

Image

Field	Type	Description
imageUrl	string	URL of the largest still image Pinterest publishes for this result. The image file itself is never copied or re-hosted.
imageWidth	integer	Width in pixels of the image at imageUrl.
imageHeight	integer	Height in pixels of the image at imageUrl.
altText	string	Alt text of the image: the one written for the pin when there is one, otherwise the one Pinterest generated.
dominantColor	string	Hex colour Pinterest computed for the image, e.g. #816748. Handy for palette and trend analysis.

Board

Field	Type	Description
boardName	string	Name of the board the pin is saved on.
boardUrl	string	Absolute URL of that board.
pinCount	integer	Number of pins on the board. Board rows only.

Engagement (complete for pin URLs, partial elsewhere — see Limitations)

Field	Type	Description
saveCount	integer	Times the pin was saved. Returned for a pin URL and for board pins; usually absent from search results.
repinCount	integer	Repin count. Returned for a pin URL and for board pins; usually absent from search results.
commentCount	integer	Comment count. Returned for a pin URL; usually absent from search results.
shareCount	integer	Times the pin was shared. Pin URLs only.
reactionCount	integer	Reactions of every kind added up. Returned for a pin URL and for most search results.

Bookkeeping

Field	Type	Description
searchQuery	string	The keyword this row was found with, so several searches stay separable in one dataset. Absent on rows that came from a board URL or a pin URL.
url	string	The URL this row describes; identical to pinUrl.
status	string	ok, not_found or error — see the table above.
error	string	Why the row is not ok. Absent on ok rows.
scrapedAt	string	ISO 8601 timestamp of when this row was extracted.

How to scrape Pinterest pins and boards

1. Open the Actor and type your keywords into Keywords, one per line (or a single one into Search keyword). To read a board or a single pin instead, paste its URL into Pinterest URLs; search URLs go there too, and a bare keyword typed into that field works as well.
2. Pick a Scope for the keyword searches: Pins for individual pins (the default), Boards to find boards about the topic.
3. Set Maximum results. Start with the default 20 to see the shape of the data. With several keywords the total is shared evenly between them; set Maximum results per search to choose the share yourself.
4. Click Start. Results appear in the dataset within seconds; export as JSON, CSV, Excel, XML or HTML, or read them from the API.

To run it from code or from an agent, call the Apify API with the same input object; the dataset is available at https://api.apify.com/v2/datasets/<datasetId>/items.

Input

Field	Type	Default	When to change it
startUrls	array of strings	["https://www.pinterest.com/search/pins/?q=coffee"]	Replace with your own URLs, one per line: search pages, boards (https://www.pinterest.com/<owner>/<board>/) or single pins (https://www.pinterest.com/pin/<id>/), in any mix. Country domains (pinterest.co.uk, pinterest.de, ...) are accepted, and a bare keyword such as coffee works as a shortcut.
query	string	(empty)	Use it instead of a URL when you just want to type one keyword the way you would in the Pinterest search box. It can be combined with startUrls: it simply adds one more search.
queries	array of strings	(empty)	Several keywords in one run, one per line. Each is searched at the selected Scope, and searchQuery on every row says which keyword found it. When queries or query is given and startUrls still holds only the built-in coffee example, the example is skipped.
scope	select: pins, boards	pins	Switch to boards when you want the boards that rank for a topic rather than individual pins. Applies to query, to queries and to any search URL that does not name a scope itself. Board and pin URLs are not affected.
maxItems	integer	20	Raise it for a real dataset — this is also your spending cap, since each ok row is one billable result. It counts rows of every status, across all keywords and URLs.
maxItemsPerQuery	integer	(empty)	The most rows any one keyword, search URL or board may add. Leave it empty and a single search or board may use everything up to maxItems; with queries, leaving it empty shares maxItems evenly between the searches so the first keyword cannot use up the run.
proxyConfiguration	object	{ "useApifyProxy": false }	Leave off; the Actor is HTTP-only and needs no proxy from a normal IP. Enable Apify Proxy (datacenter) only if Pinterest starts answering 403 or 429.

Example input — several keywords in one run, 50 pins each:

{
    "queries": ["cold brew", "matcha latte", "espresso bar"],
    "scope": "pins",
    "maxItems": 150,
    "maxItemsPerQuery": 50
}

Example input — every pin saved on a board:

{
    "startUrls": ["https://www.pinterest.com/ikeahacker/best-ikea-hacks/"],
    "maxItems": 200
}

Example input — single pins in detail, with counts and the video URL:

{
    "startUrls": [
        "https://www.pinterest.com/pin/181199585001797964/",
        "https://www.pinterest.com/pin/63191201017606220/"
    ]
}

The original forms keep working unchanged: a search URL in startUrls, a single keyword in query, or both.

{
    "startUrls": ["https://www.pinterest.com/search/pins/?q=coffee"],
    "query": "cold brew setup",
    "scope": "pins",
    "maxItems": 100,
    "proxyConfiguration": { "useApifyProxy": false }
}

Profile URLs, board section URLs and /search/users/ pages are not supported. They are skipped, each with one free error row explaining why, rather than silently disappearing. Profiles and /search/users/ are refused on purpose: this Actor does not collect people.

Output

A real row from the default run (start URL https://www.pinterest.com/search/pins/?q=coffee, maxItems: 20, which finished in 6.3 seconds with 20 ok rows from a single HTTP request):

{
    "url": "https://www.pinterest.com/pin/26036504092933313/",
    "status": "ok",
    "scrapedAt": "2026-09-20T19:10:30.981Z",
    "resultType": "pin",
    "pinId": "26036504092933313",
    "searchQuery": "coffee",
    "title": "there is a cup of coffee on the saucer",
    "pinUrl": "https://www.pinterest.com/pin/26036504092933313/",
    "imageUrl": "https://i.pinimg.com/originals/f3/56/59/f3565938a74dffad34103837416f664f.jpg",
    "imageWidth": 1440,
    "imageHeight": 1440,
    "altText": "there is a cup of coffee on the saucer",
    "dominantColor": "#7c6e5d",
    "boardName": "Ideias para meu diário",
    "boardUrl": "https://www.pinterest.com/giovannaemilia2/ideias-para-meu-di%C3%A1rio/",
    "createdAt": "2026-01-17T20:05:53.000Z",
    "reactionCount": 2641,
    "isVideo": false
}

A pin URL (https://www.pinterest.com/pin/181199585001797964/), a video pin, from a real run:

{
    "url": "https://www.pinterest.com/pin/181199585001797964/",
    "status": "ok",
    "scrapedAt": "2026-09-20T19:07:41.040Z",
    "resultType": "pin",
    "pinId": "181199585001797964",
    "title": "Pouring Perfection: Tulip Latte Art in Action",
    "description": "Perfecting the tulip pour, one latte at a time 🌷☕️ #LatteGoals",
    "link": "https://subminimal.com/products/nanofoamer",
    "domain": "subminimal.com",
    "pinUrl": "https://www.pinterest.com/pin/181199585001797964/",
    "imageUrl": "https://i.pinimg.com/originals/9e/15/f9/9e15f947032f2dfa5f0d119332a47e8a.jpg",
    "imageWidth": 1080,
    "imageHeight": 1920,
    "dominantColor": "#b27755",
    "boardName": "Coffee is Life",
    "boardUrl": "https://www.pinterest.com/Kathyd2016/coffee-is-life/",
    "createdAt": "2024-06-20T10:33:25.000Z",
    "saveCount": 1423,
    "repinCount": 582,
    "commentCount": 0,
    "shareCount": 54,
    "reactionCount": 194,
    "isVideo": true,
    "videoUrl": "https://v1.pinimg.com/videos/iht/720p/0b/08/e9/0b08e9d5e378b9f8f7f3081de3bfd394.mp4"
}

One pin of a board URL (https://www.pinterest.com/ikeahacker/best-ikea-hacks/), from the same run:

{
    "url": "https://www.pinterest.com/pin/112167846967490473/",
    "status": "ok",
    "scrapedAt": "2026-09-20T19:07:41.594Z",
    "resultType": "pin",
    "pinId": "112167846967490473",
    "title": "Turn IKEA doll beds into an adorable cat bunk bed",
    "description": "Move over dolls, we want the IKEA DUKTIG as a cat bunk bed You may have seen some photos of these adorable cat bunk beds and wondered where they came from. They’re actually sold as “doll beds” at IKEA! IKEA does sell pet furniture (in some countries) and it’s probably easier to buy than make,",
    "link": "https://ikeahackers.net/2019/10/cat-bunk-bed-pets-ikea.html",
    "domain": "ikeahackers.net",
    "pinUrl": "https://www.pinterest.com/pin/112167846967490473/",
    "imageUrl": "https://i.pinimg.com/originals/82/60/9e/82609efa4694110d07ee9ec0cf77856e.jpg",
    "imageWidth": 1000,
    "imageHeight": 1500,
    "altText": "there are two pictures of a cat on the floor and in front of a bunk bed",
    "dominantColor": "#ab9e8d",
    "boardName": "Best Ikea Hacks",
    "boardUrl": "https://www.pinterest.com/ikeahacker/best-ikea-hacks/",
    "saveCount": 1,
    "repinCount": 0,
    "isVideo": false
}

The same Actor with "scope": "boards" and "query": "minimalist home office":

{
    "url": "https://www.pinterest.com/greenwithless/minimalist-home-office/",
    "status": "ok",
    "scrapedAt": "2026-09-08T12:48:20.186Z",
    "resultType": "board",
    "pinId": "848084242269709151",
    "searchQuery": "minimalist home office",
    "title": "Minimalist Home Office",
    "description": "Do you like minimalist home office design? On this board, you'll find the best minimalist home office ideas, minimalist home office decor, minimalist home office setup ideas, minimalist desk setup ideas, decluttering ideas for a minimalist home office, and more!",
    "pinUrl": "https://www.pinterest.com/greenwithless/minimalist-home-office/",
    "imageUrl": "https://i.pinimg.com/474x/77/23/b5/7723b511a95d32109589011837e0dfbe.jpg",
    "boardName": "Minimalist Home Office",
    "boardUrl": "https://www.pinterest.com/greenwithless/minimalist-home-office/",
    "pinCount": 29,
    "isVideo": false
}

And a profile URL, refused and written as a free row instead of vanishing:

{
    "url": "https://www.pinterest.com/",
    "status": "error",
    "error": "Input entry \"Profile URLs are not supported (this Actor does not collect people): https://www.pinterest.com/ikeahacker/\" is not a URL this Actor can fetch; it was skipped.",
    "scrapedAt": "2026-09-20T19:07:39.672Z"
}

How much does it cost to scrape Pinterest?

Pay per event. You pay for what you actually receive.

Event	Price	When it is charged
Actor start	$0.005	Once per run (per GB of memory; the 1 GB default is one event).
Result	$0.00199	Per saved row with status ok — one pin (from a search, a board or a pin URL), or one board when Scope is boards.

That works out at $1.99 per 1,000 results. Rows with status not_found or error are written to your dataset free of charge, and the Actor never bills for a row it could not deliver. Platform compute is included: the two events above are the whole price list. The free Apify plan's $5 of monthly credit covers about 2,500 results in a single run.

How it compares

This Actor is deliberately focused and cheap: keyword search across pins and boards, the pins of a board, and single pins in detail — all over HTTP in seconds, with no login, no cookie and no browser to pay for. If what you want is a wide sample of what Pinterest surfaces for a topic, with real image URLs, destination links and colour data you can analyse, it is a good fit and it will not cost much.

It is not the tool for creator analytics or influencer discovery: it returns no pinner identity data at all, it does not read profiles, and full engagement counts come only with pin URLs — a search response rarely carries them. It cannot read private boards, it does not download image files, and because Pinterest personalises and localises search, two runs of the same keyword will not return an identical list. If you need per-creator metrics, follower counts or reproducible ranking, look elsewhere.

Integrations, API and AI agents

Results export to CSV, JSON, Excel and XML from the Apify console, or flow straight into Google Sheets, Airtable, Slack or a CRM through Apify integrations and webhooks — including Make, Zapier and n8n. Schedule the Actor to re-run a set of keywords weekly and only append what is new.

It can be called from the Apify API, from any Apify client library, from the Apify MCP server, or by any AI agent that can call an HTTP endpoint. Output is flat JSON with a stable status column, so an agent can distinguish "found nothing" from "could not check" without heuristics.

Tips

• Use searchQuery when you batch. Put twenty keywords into queries and every row still tells you which keyword produced it.
• Two steps for full counts and MP4s. Search first, then feed the pinUrl values you care about back in as startUrls: each comes back with save, repin, share and reaction counts and, for a video pin, the MP4 instead of the HLS playlist. One pin URL is one request and one result.
• dominantColor is the cheap trend signal. Bucket it by hue and plot against createdAt to see a palette shift over months without downloading a single image.
• link is the commercial signal. Filter to rows where link is present to get only pins that actually send traffic somewhere.
• Raise maxItems before you raise concurrency. One request already returns up to 100 results, so a bigger maxItems is nearly free in requests.
• Set maxItems deliberately. It counts every row, including free not_found and error rows, so a run always stays bounded.

Limitations

• Unofficial endpoints. This Actor reads the undocumented JSON calls that pinterest.com's own search, board and pin pages make. They are not a published API, they carry no stability guarantee, and Pinterest can change or restrict them at any time. If that happens the Actor fails loudly with error rows rather than returning an empty success.
• No pinner identity, by design. Usernames, real names, profile URLs, avatars and follower counts are present in the upstream response and are never written to the dataset. If you need creator-level data, this Actor is not the right tool. Note that a boardUrl naturally contains the board owner's public handle as its first path segment, because that is how Pinterest builds board URLs.
• Engagement counts depend on the path. A pin URL returns saves, repins, comments, shares and reactions. A board listing returns saves and repins. A search response usually carries only reactionCount. Treat counts on search rows as a bonus, not as a guarantee; for dependable counts, pass the pin URLs.
• Video URLs depend on the path too. A pin URL returns an MP4. Search results and board listings carry only the HLS playlist (.m3u8), which players and ffmpeg read but a plain download does not. Video files are referenced, never downloaded.
• Board listings have no creation date. Pinterest leaves createdAt out of a board's pin list; pass a pin URL when you need it.
• Board sections and short links are not read. A section URL (/<owner>/<board>/<section>/) and pin.it short links are skipped with an explanatory row; use the board URL or the full pin URL. Secret boards are not reachable.
• Results vary by region and over time. Pinterest personalises and localises search. The same keyword run from a different country, or an hour later, returns a different and differently ordered set. Row counts and ordering are therefore not reproducible.
• description and link are often empty. Most pins carry neither. In a typical 100-pin sample roughly a third have a description and under a fifth have an outbound link. title falls back to Pinterest's alt text and is present on almost every row.
• A pins search almost never returns nothing. For an unmatched keyword Pinterest quietly serves loosely related pins instead of an empty result, so a not_found row is rare in pins scope. A boards search does return a genuinely empty result set, and that becomes one not_found row.
• Images are referenced, never downloaded. You get imageUrl; fetching or storing the image itself is your decision and your responsibility.
• Datacenter IPs are untested at scale. The endpoint was verified working without a proxy, but Pinterest may treat cloud IP ranges differently from a home connection. If you see HTTP 403 or 429 the run will fail visibly — switch on Apify Proxy (datacenter) in the input and retry.

FAQ

Is it legal to scrape Pinterest? This Actor reads only publicly visible search results, without logging in and without defeating any access control. Whether a particular collection and use is lawful depends on your jurisdiction, your purpose and Pinterest's terms of service, which you are responsible for reading. Get your own legal advice before running this at scale or for a commercial product.

Do I need a Pinterest account, an API key or a proxy? None of the three. The Actor reads a public search response over plain HTTP from Apify's datacenter IPs. Apify Proxy is available in the input but is off by default and is only worth enabling if you start seeing 403 or 429 responses.

Can I scrape Pinterest boards as well as pins? Yes, in two ways. To find boards about a topic, set Scope to Boards and each row becomes a board: title is the board name, pinCount is how many pins it holds, and pinUrl links to it. To get the pins of a board, paste the board URL into startUrls.

Can I get the video of a video pin? Yes. videoUrl holds the best URL Pinterest publishes: an MP4 when you pass the pin URL, the HLS playlist on search and board rows. The Actor never downloads the file.

Can I search several keywords at once? Yes. Put them into queries, one per line. maxItems is shared evenly between them unless you set maxItemsPerQuery.

How many pins can I get in one run? As many as maxItems allows. One request to Pinterest returns 25-100 results, so even a 1,000-row run is only 10-40 requests and finishes in well under a minute.

Why did I get fewer rows than maxItems? Either the search genuinely ran out of results, or maxItems counts rows of every status — a couple of free error rows for unusable input entries reduce the number of ok rows in the same run. The final log line reports the exact breakdown.

What if a field is missing? Missing means Pinterest did not publish it for that result. Fields are omitted rather than filled with empty strings or zeros, so a missing link means "this pin points nowhere", not "the scraper failed". Only error rows indicate a failure.

Can I search for people, or scrape a profile? No. Board search is supported through the Scope input and a single board can be read by its URL; people search and profile pages are deliberately not, and /search/users/ and profile URLs are rejected with an explanatory row. This Actor is for content, not for profiling individuals.

Does it download the images? No. It gives you imageUrl, imageWidth and imageHeight. No image file is copied, cached or re-hosted by the Actor.

Can I use it from an AI agent or an MCP client? Yes. Call it through the Apify API or the Apify MCP server like any other Actor; the flat JSON rows and the status column are designed for exactly that.

Legal and data-protection notice

This Actor collects only publicly accessible content from Pinterest search results, public boards and public pins. It does not log in, does not access private boards, and does not attempt to defeat any technical access restriction.

Pinterest content is user-generated, which means personal data can appear inside it. This Actor is built to avoid that: pinner usernames, full names, profile URLs, profile images and follower counts are present in the upstream response and are deliberately excluded from every dataset row. Board names and board URLs are retained as content metadata, and a board URL includes the owner's public handle because that is part of the URL Pinterest publishes. Images and videos are referenced by URL only; no media file is copied or stored by the Actor.

If your dataset nevertheless ends up containing personal data — for example a person's name written into a pin title, a description or a board name by the person who posted it — you are the data controller for that processing under the GDPR, the CCPA and comparable laws. You are responsible for having a lawful basis, for honouring data-subject requests, and for deleting data you no longer need. Do not use this Actor to build profiles of individuals.

Pinterest is a trademark of its respective owner. This Actor is an independent, unofficial tool and is not affiliated with, authorised by or endorsed by Pinterest.