Facebook Public Data API — Try the interactive demo

Pass a Facebook Page URL in url. The observed response has a results object with name, type, page_id, url, verified, followers, categories, image, cover_image, delegate_page.id and reels_page_id. Some optional contact and biography fields were null; do not assume they are populated. Use page_id for Page posts and retain the distinct delegate and Reels identifiers for their respective routes. A nonexistent URL returned results.type=private_page with no identity: this response cannot distinguish an absent Page from a private or inaccessible one and is not a reliable existence check. Follower counts are public observations, not private reach or Page Insights.

Supply page_id from page-details and omit cursor for the first request. The observed body contains results and a cursor string. Posts include post_id, url, message, message_rich, numeric timestamp, author, authors, comments_count, reactions_count, reshare_count, reaction-category counts and nullable media fields such as image, video and video_files. comments_id and shares_id were also present; preserve these endpoint-specific identifiers rather than assuming every downstream route uses post_id. Treat top_comments as a preview, not the whole discussion. Follow returned cursors unchanged with a request bound and ID deduplication; complete history is not guaranteed. start_date and end_date are documented inputs, but their exact formats and behavior have not been validated.

Enter query to discover Page candidates; location_uid and cursor are optional documented inputs. A NASA query returned results with name, facebook_id, url, profile_url, type, is_verified and an image object containing uri and dimensions, plus a cursor string. Verify candidate identity before using a search result as the intended organization. A deliberately nonexistent query returned exactly results=[] and cursor=null: an empty search is a valid outcome, not automatically an outage. Reuse a returned non-null cursor unchanged; location filtering and continuation were not exercised in the samples.

Provide url for a Facebook Page to request its Page identifier. This is a documented URL-to-ID resolver; no representative response for this route was sampled, so no response envelope or exact field layout is asserted. Confirm the returned identifier before passing it to Page posts, reviews or other ID-based reads. The separately sampled page-details operation is an alternative when a name, verification flag and follower observation are also needed. Resolving an identifier does not establish access to every Page content surface.

Supply page_id to request recent publicly accessible Page photos. For continuation, the documented contract requires both cursor and collection_id from the first response; keep the pair together for the same Page. This route has not been sampled, so photo fields, result envelope and ordering details are unverified. Inspect a representative response before defining a storage schema. Photo access does not grant download, redistribution or model-training rights, and cursor traversal does not promise a complete photo archive.

Use reels_page_id obtained from Page details when available. The documented page_id input is legacy and should not be substituted for the Reels collection identifier. Both inputs remain optional in the parameter schema for compatibility, but a usable collection identifier is a practical prerequisite for the recommended flow. cursor is available for continuation. No Page Reels response was sampled; do not assume the Page-post envelope or fields apply. Reels availability and historical coverage depend on public visibility, and viewing references conveys no media reuse rights.

Pass delegate_page_id, not page_id. The sampled Page details returned delegate_page.id, and Page-post authors also included a delegate identifier; use the value for the intended Page and preserve its distinct meaning. cursor is a documented continuation input. This video route has not been sampled, so no specific response keys, playable formats or video counts are promised. Keep returned references within source visibility and media permissions; a delegate identifier does not unlock private videos.

Supply a business page_id and optionally a cursor from a prior review response. This documented route requests public Page reviews, but no review response was sampled. Validate the response layout and any review text, dates or ratings actually supplied before designing analysis. Availability can vary by Page and visibility; an empty or unavailable response is not proof that a business has no reviews. Client-managed snapshots and human-led theme analysis are possible workflows, not built-in monitoring or sentiment features.

Provide page_id to request the documented Page transparency record. This endpoint has not been sampled, so specific provenance, management, location or change-history fields are not guaranteed. Inspect what is actually returned and retain an observation timestamp before using it in a due-diligence workflow. Transparency information is context for human review, not a verified identity certificate, automated risk score or guarantee of authenticity.

Supply query to search public posts; cursor, location_uid, recent_posts, start_date and end_date are documented optional inputs. The sampled NASA search returned results and a cursor, with post_id, url, message, numeric timestamp, author identity, comments_count, reactions_count, reshare_count, reaction breakdowns and nullable image or video references. Results included an official NASA post and related reporting, so relevance is not an exact-author filter. Only the query-only request was exercised; date formats, location filtering, recent_posts and continuation behavior still need validation. Search results are observations, not an exhaustive count of mentions or historical coverage.

Supply post_id or post_url; when both are supplied, post_id takes precedence. Omit the unused ID when looking up a URL so it cannot override the intended input. This is a documented post-detail route with no sampled response, so no exact envelope or detail fields are asserted. Validate any returned comment and share identifiers before chaining discussion or reshare reads; those identifiers need not equal the visible post ID. Access is limited by public availability, and an unavailable result alone does not prove deletion.

Provide the required post_id and an optional cursor from the same discussion. The documented discussion workflow refers to a comment-collection identifier supplied with post data, so preserve comments_id when present and validate which identifier this route accepts before bulk traversal; do not assume every post ID form is interchangeable. No response for this endpoint was sampled, so comment text, author, timestamps and result shape are not asserted here. A top_comments preview from Page posts is not a full comment collection. Respect the privacy of individuals and use bounded, permission-aware samples.

Use the legacy comment identifier as comment_id and preserve the matching post_id and expansion_token when the discussion response supplies them. The parameter schema lists these inputs as optional, but it does not establish that an empty request can resolve a thread. Treat expansion_token as a sensitive opaque continuation value: obtain it for the intended thread and do not publish it in shared examples. This endpoint has not been sampled, so parent-child fields, reply completeness and the precise input combination must be validated. Do not substitute a newer opaque comment identifier for the documented legacy ID.

Pass a known public comment_id to request one comment directly. This documented endpoint has no sampled response; text, author, creation time and exact result shape must be checked before use. It is distinct from the nested-reply route and does not promise a reply collection. Refresh only comments relevant to a permitted review purpose; a missing or inaccessible record must be treated as unavailable, not conclusively deleted.

Supply post_id, with optional reaction_type and cursor. The documented reaction choices are like, love, haha, wow, sad and care; select one supported value rather than sending the comma-separated list of choices. This route has not been sampled, so reaction-record fields and filter behavior are unverified. It is separate from the aggregated reaction counts observed in Page-post and search-post responses. Visible reactions are neither complete audience reach nor a built-in sentiment classification.

Pass the share identifier supplied with post details in the public parameter named post_id; the documented prerequisite calls it share_id, while the sampled Page-post feed used shares_id. Preserve the exact field and value from the source response and validate compatibility for this route rather than assuming all post identifiers are equal. cursor is optional. No reshare response was sampled, so its field layout and coverage are unverified. Only visible reshares may be accessible; this does not expose private sharing or a complete dissemination network.

Provide query, a two-letter country code and one active_status value: ALL, ACTIVE or INACTIVE. The observed ACTIVE, US, pizza request returned an ads array and a cursor. Ad records include ad_archive_id, page_id, page_name, is_active, publisher_platform, start_date, end_date and snapshot data such as creative cards, text, links, titles and media references. Dynamic-ad template strings can coexist with populated snapshot.cards; preserve them as templates rather than dismissing the record as mock data. The other status choices and continuation were not exercised. Public ad discovery is not official Meta API access, campaign management or a guarantee of spend, impressions or complete coverage.

Supply page_id, a two-letter country and one active_status value from ALL, ACTIVE or INACTIVE. cursor and the boolean sort_by_new are optional documented inputs. This advertiser-specific route was not sampled; do not assume the search-ads payload is identical or that sort_by_new guarantees a particular chronology. Inspect returned archive identifiers and creative fields before joining records. Repeated checks and new-ad detection require your own scheduling and storage; this read does not create alerts or manage campaigns.

Provide the matching ad_archive_id, page_id and two-letter country for a public advertisement. The sampled search-ads response supplies archive and advertiser IDs, but this detail endpoint itself has not been sampled. Validate its exact response layout and available creative fields before using it for enrichment. Keep the three inputs associated with the same advertisement; a public ad lookup does not grant campaign-editing access or guarantee spend and audience statistics.

Pass ad_archive_id when only a saved public ad archive identifier is available. Documentation warns that this convenience route can be less reliable; use get-ad-details with the matching country and Page ID when available. No response for this endpoint was sampled, so neither creative fields nor an equivalence with the contextual detail route is guaranteed. A failure here alone must not be interpreted as evidence that an ad was deleted.

Provide query and country to request country-aware ad-search suggestions. This documented autocomplete route has not been sampled, so suggestion text, advertiser keys, ordering and response shape must be validated. Use suggestions to refine a discovery query, then deliberately run the relevant ad search. Autocomplete is not an advertising-targeting audience, proof of advertiser identity or a complete list of brands in a market.

Pass a public profile URL in url. The observed Mark Zuckerberg response has a profile object with name, profile_id, url, verified, intro, image, cover_image, about_public, influencer_category and Reels collection identifiers. The about object was empty and delegate_page_id was null; do not promise these fields will be populated. Other public biography attributes may be present but should be minimized for the intended permitted purpose. This read does not expose private profile fields or prove identity for other similarly named people. Obtain appropriate permission for creator research and avoid inferring sensitive traits.

Supply a known profile_id to request public profile details without resolving a URL first. This route has not been sampled; the separately observed URL-based profile response must not be assumed to define its exact envelope or fields. Validate representative output and minimize any personal data retained. A stable identifier helps join authorized observations but is not an identity-verification certificate, and it does not grant access to private profile information.

Provide a public Facebook profile url to request its identifier. The resolver is documented but was not sampled, so no exact return key or envelope is guaranteed. Validate the identifier before using ID-based profile or content routes. The sampled profile-details-by-url route is an alternative when you also need the visible name and profile context. URL normalization should preserve the original submitted link for human identity checks; it does not establish consent or access to non-public content.

Supply profile_id and optionally cursor, start_date or end_date. This documented route requests recent posts when public, but it has not been sampled; neither the Page-post envelope nor its detailed fields are guaranteed here. Date input formats and behavior also need validation. Use only public-figure content or an appropriate, permission-aware creator workflow, and stop bounded traversal on an absent or repeated continuation value. Historical completeness and non-public posts are not promised.

Provide profile_id to request publicly visible profile photos. Optional cursor and collection_id inputs are documented; preserve any returned collection context when assessing continuation. This route has no sampled response, so exact photo fields and whether collection_id is needed for a particular next-page request must be validated. Limit use to a permitted public or creator-authorized purpose. Public visibility does not grant image reuse rights or access to private albums.

Use reels_profile_id obtained from profile-details-by-url or a validated profile-details response, with optional cursor for continuation. This is a collection identifier, not the ordinary profile_id. The successful URL-profile sample supplied such an identifier, but this Reels endpoint itself was not sampled; result fields, media formats and collection coverage remain unverified. Use public-figure or consenting-creator content within applicable permissions and maintain a Reels-specific continuation chain.

Supply profile_id for the documented visible-following lookup. To continue, pass both cursor and collection_id from the response for that same profile. This route has not been sampled, so followed-account fields, counts and completeness are unverified. It does not imply a private friend list, hidden relationships or permission to profile individuals. Use a consenting creator or another appropriate permitted purpose and minimize retained account references.

Provide query for documented global content discovery, with optional cursor, location_uid, recent_posts, start_date and end_date. No global-search response was sampled; result categories, envelope, field layout and optional-filter behavior are unverified. Use the response as an exploratory candidate source, then validate content-specific routes for records you select. Date formats must be checked rather than assumed. This is not a comprehensive platform-wide mention index or a guarantee of private-content access.

Provide hashtag and optionally a cursor from the same hashtag search. The documented route requests hashtag-related content but was not sampled, so result fields, matching rules and ranking are unverified. Validate the accepted hashtag syntax and actual output before presenting a production search workflow. Returned observations do not establish total hashtag volume, reach, sentiment or exhaustive campaign coverage.

Provide query and optionally cursor to request documented people-search candidates. This endpoint has not been sampled, so profile fields, result shape and identity matching are unverified. Treat every result as a candidate for human confirmation, not proof that a particular person owns an account. Use public-figure or opted-in creator discovery within appropriate permissions; no private contact details, sensitive-attribute inference or identity verification is promised.

Supply query and optionally location_uid or cursor for documented place discovery. No response for this route was sampled; venue identities, geographic fields and result shape must be validated. Use locations to evaluate location identifiers when needed, and confirm ambiguous venue matches manually. This route finds candidate public place references, not verified business ownership, accurate opening hours or a guarantee of geospatial completeness.

Enter query, with optional cursor, recent_videos, start_date and end_date. This documented video-search route was not sampled, so video fields, ranking and filter effects are unverified. Check the actual result structure and accepted date format before building a video index. The boolean recency input is not a freshness guarantee, and returned references do not imply streaming, downloading or reuse rights.

Supply a place-name query to request documented Facebook location identifiers. If matches are ambiguous, include the country or try an alternate spelling without diacritics. No location-search response was sampled, so exact identifier fields and geographic granularity need validation. Confirm the intended location before using its value as location_uid in supported searches; a text match is not a coordinate-based radius guarantee.

Provide query and optionally cursor for documented group discovery. This route has not been sampled, so group names, identifiers, visibility flags and exact result shape are not guaranteed. A search match does not establish that its posts or media are publicly readable; inspect accessibility separately before any content request. Use candidate discovery for appropriate community research, not private-group access or a member directory.

Pass url for a Facebook group to request its documented summary. This endpoint was not sampled, so specific summary fields, group identifiers and visibility indicators must be validated. A summary is not a guarantee that posts, photos, videos or event content can be read. Check public accessibility and appropriate permissions before selecting a group for further analysis; private-group access and member enumeration are outside the stated scope.

Supply a group url to request its documented Facebook identifier. No response for this resolver was sampled; verify the returned field and identifier before passing it as group_id to content routes. A resolved ID normalizes a reference but does not prove that the group is public or that its content is available. Preserve the original URL when reconciling group references and avoid treating identity resolution as access authorization.

Provide group_id, optionally cursor and one sorting_order value: CHRONOLOGICAL or TOP_POSTS. Documentation limits this route to public groups; an empty or unavailable response warrants checking public visibility rather than assuming no discussions exist. No group-post response was sampled, so post fields, ordering behavior and result envelope remain unverified. Use bounded sampling, maintain observation times and validate IDs before deduplication. The sorting controls do not promise a complete historical archive.

Supply query and the intended group_id for a group-scoped post search. The schema marks group_id optional, but omitting it does not establish a useful group-specific workflow; provide a confirmed public-group identifier rather than relying on a saved default. cursor, start_date and end_date are optional documented inputs. This route has not been sampled, so matching, field layout and date formats need validation. A query does not enable access to private-group discussions or prove complete topic coverage.

Supply group_id and optionally a cursor from the same public-group photo read. This documented route is limited to publicly accessible groups and was not sampled; photo fields, collection ordering and media behavior are unverified. Confirm public access if no content is returned. Store only references needed for a permitted visual-review purpose, deduplicate validated identifiers and do not treat the result as a full group archive or a license to reuse images.

Pass group_id for a publicly accessible group and optionally cursor to continue that video collection. The endpoint is documented but not sampled, so video fields, media formats and result envelope need validation. Verify group public visibility when content is unavailable. Use the returned material only for an appropriate permitted purpose; neither private-group access nor download and reuse rights are implied.

Supply group_id for the intended public group and optionally cursor from its event response. The documented route requests future events, but no response was sampled; event identifiers, dates and exact shape need validation. An empty response may reflect public visibility or availability, not necessarily the absence of planned activity. Use a bounded, human-reviewed collection process, and do not infer private group calendars or exhaustive event coverage.

Provide query, with optional location_uid, cursor, start_date and end_date. This documented event-search endpoint has not been sampled, so event fields, geographic matching and date-filter formats are unverified. Resolve and confirm location identifiers separately when needed. Treat returned records as candidates for human review before adding them to a local calendar; neither event completeness nor current schedules are guaranteed.

Pass url for a publicly accessible Facebook event. This is a documented event-detail lookup with no sampled response, so title, venue, dates and result envelope are not asserted. Validate returned fields before importing them into a calendar, and preserve the source URL for editorial checks. Event availability or scheduling may change, and a readable link is not permission to republish unrelated personal attendee data.

Supply a confirmed event_id to request the documented event-detail record. This route was not sampled, so exact fields and equivalence with URL-based event lookup remain unverified. Use only identifiers reconciled from actual event references, and retain observation times when refreshing stored records. An unavailable response does not establish cancellation; validate the returned information or check the public source before changing a published calendar.

Provide page_id and optionally cursor and collection_id from a Page event response. This documented route requests future events associated with a Page, but it was not sampled; exact event fields and pagination requirements must be validated. Preserve any returned collection context rather than inventing it. Use the result as a reviewed public-event feed, not a guarantee of all future activity or a live venue schedule.

Supply page_id with optional cursor and collection_id for the documented past-event collection. No response was sampled, so event fields, ordering and continuation details need validation. Retain any returned collection identifier with its matching cursor. Use this as a source of observed public past-event references; pagination is not a promise of complete history or evidence that every returned event actually occurred as planned.

Supply query, lat and lng, with optional radius_km, category, condition, price_min and price_max. Documentation gives category examples cars and boats, and condition values new, used_like_new, used_good and used_fair; conditions may be combined with commas. No Marketplace response was sampled, so listing fields, distance behavior and price interpretation must be validated. The exposed endpoint schema has no cursor input despite a general pagination mention, so this integration does not invent one or promise pagination. Treat any returned listing prices as asking-price observations, not completed-sale values or private seller contacts.

Supply a known listing_id from a validated public Marketplace reference. This documented detail route was not sampled, so description, price, location, seller and availability fields are not guaranteed. Inspect the actual response before adding records to a comparison dataset. A listing disappearing from a public lookup does not prove a sale, and a displayed price is not evidence of the final transaction value. No private seller-contact access is promised.

Provide query and choose one sort_order value, VIEWERS or SUGGESTED; cursor is optional. Documentation also describes an empty-query browse mode, but query is marked required, so provide a search term until empty-input behavior is validated. No gaming-directory response was sampled; game fields, ordering and browse behavior need testing. Documented viewer-based sorting is not a verified platform-wide popularity measure.

Supply a confirmed game_id and one sort_order value, VIEWERS or SUGGESTED, with optional cursor. The documented route requests visible game-related live videos, but no response was sampled; broadcast fields, live-state accuracy and ranking remain unverified. Validate game identifiers from actual game discovery before chaining. Returned references are for permission-aware discovery, not a guarantee that a stream remains live or a right to rebroadcast it.