Facebook Photos Scraper

Collect photo galleries from public Facebook Pages into one structured dataset. Supply Page URLs or usernames and receive a row for each unique photo, with its image URL, Facebook permalink, thumbnail, dimensions and available accessibility text.

This Actor is a private development candidate. Public release and human source/data-use review remain pending. The examples below describe the implemented input and output contract; the sample output is illustrative.

Input

{
    "startUrls": ["https://www.facebook.com/NASA"],
    "maxPhotos": 5,
    "maxPhotosPerPage": 5,
    "maxRequests": 2,
    "maxProviderCostUsd": 0.01
}

Field	Default	Behavior
startUrls	Required	1–20 public Page roots, /photos URLs, or usernames.
maxPhotos	100	Global maximum of unique saved photos, from 1 to 10,000.
maxPhotosPerPage	50	Maximum saved photos per Facebook Page, from 1 to 10,000.
maxRequests	30	Global maximum upstream requests, from 1 to 100. Empty and failed attempts count.
maxProviderCostUsd	0.10	Upstream cost reservation cap, from $0.001 to $0.50. This excludes Apify usage and event fees.

Both photo limits apply. The request and provider-cost limits can stop a run before it reaches the requested number of photos. A Page with fewer accessible photos returns fewer rows. Repeated inputs are collapsed and photo IDs are deduplicated across the run. Requests are serial and have no automatic retries.

Output

{
    "photoId": "1001",
    "photoUrl": "https://www.facebook.com/photo.php?fbid=1001",
    "sourceUrl": "https://www.facebook.com/examplepage",
    "imageUrl": "https://scontent.xx.fbcdn.net/illustrative-image.jpg",
    "thumbnailUrl": null,
    "width": 1080,
    "height": 720,
    "altText": "May be an image of a mountain",
    "provider": "scrapecreators",
    "scrapedAt": "2026-10-26T00:00:00.000Z"
}

imageUrl is the largest viewer image supplied by the source; original upload resolution is not guaranteed. Signed CDN URLs may expire. Missing optional fields are null. altText is the accessibility caption supplied by Facebook. scrapedAt is the collection time.

Use the Dataset tab to export JSON, CSV or Excel. The output schema exposes the dataset, OUTPUT and RUN_SUMMARY for API and agent consumers. After deployment, Apify's hosted MCP can run the batch Actor with the same input; a hosted MCP call has not yet been tested.

Scope and limits

The Actor collects accessible public Page galleries without Facebook credentials. It does not accept cookies or access private/friends-only content. Personal-profile gallery coverage is unverified; a numeric profile.php?id=... Page URL is accepted but does not guarantee personal-profile access.

Individual photo permalinks, album crawling, tagged-photo selection, videos, dates, post captions, reactions and image-file downloads are outside this version's contract. For individual media URLs, use Facebook Media Downloader. For Page metadata, use Facebook Page Scraper.

The Actor uses an owner-managed upstream key. Users do not enter provider secrets. Provider outages, login walls and access restrictions can reduce coverage. No secondary provider with verified equivalent gallery coverage is configured.

Charges and stop behavior

Pay per event charges are $0.005 per startup, $0.02 per gallery request, and $0.001 per saved photo ($1.00 per 1,000 photos, plus request fees, startup and platform usage). Each pagination request is a separate gallery request. For example, one request saving 8 photos costs $0.033 in events; two requests saving one photo each cost $0.047 in events. These rates apply equally across paid tiers.

The platform startup event can apply even to invalid, empty or failed runs (one event per GB of Actor memory, minimum one). The gallery request fee is accepted before an upstream attempt and applies to empty responses and failed attempts as well as successful pages. An interruption after an accepted request fee can require owner reconciliation before dispatch; it is never blindly replayed. A saved photo incurs one result event; duplicates and rejected records incur no photo fee. Before dispatch, the run reserves the gallery request fee plus room for at least one saved photo within the event limit. The event limit excludes additional platform usage. Provider spending is separately capped by the input limits.

The provider reservation uses the verified published baseline of $0.00188 per request. It is a conservative estimate, not a provider invoice. Tariff drift or a provider reporting more than the expected one credit stops work; a changed tariff can already have charged for that request. Free/non-paying traffic and failed requests can still create owner costs.

OUTPUT.outcome and RUN_SUMMARY.outcome explain the result:

Outcome	Meaning
COMPLETE	The available gallery or requested photo cap was delivered.
PARTIAL	Useful output was retained with failures, invalid rows, cancellation or a request/spending/deadline stop; caps can also stop before any row.
VALID_EMPTY	The provider successfully returned an empty gallery.
INVALID_INPUT	Correct the documented input fields.
UPSTREAM_FAILED	No useful photos were available because extraction failed. The cloud run fails.
CONFIG_ERROR	Required owner configuration or reconciliation is missing. With no confirmed output, the cloud run fails.

Pagination, duplicate IDs, reservations and pending effects are checkpointed within the same run. Confirmed writes are skipped on resurrection. An interrupted request or dataset/charge operation with an uncertain result blocks further billable work until the owner reconciles it. This is not an exactly-once or automatic recovery guarantee. Export the dataset directly if a saved row has an unresolved charge receipt.

Responsible use

Use only data you are authorized to collect and use. Public accessibility does not grant image reuse rights. This Actor is independently developed and is not affiliated with Meta or Facebook. Store publication and source/data-use review remain separate from technical testing.