Quora Question, Answer & Relationship Dataset Builder

Build a structured Quora dataset from direct links or best-effort search discovery. Point it at question, answer, profile, Topic, Space, or Post URLs and it returns normalized rows for every entity it finds, with the question → answer → comment chain and author/topic links kept intact. Choose how deep to go, how the rows are shaped, and which relationship or feed data to include.

This is an independent, third-party actor. It is not affiliated with, endorsed by, or licensed by Quora, Inc. No Quora account, login, session cookie, or paid proxy is used or required.

What this actor does

• Accepts question, answer, profile, Topic, Space, and Post URLs directly, plus best-effort search discovery. Typed seed lists (questionUrls, answerUrls, topicUrls, spaceUrls, profileUrls, postUrls) override route auto-detection.
• Empty fields are omitted; a missing field means the source did not expose it, never null or a sentinel.
• Parses Quora's server-rendered Relay payloads and double-decodes the window.ansFrontendGlobals.data.inlineQueryResults .push() strings; it does not depend on CSS classes or DOM selectors for the core fields.
• Paginates question answers with Quora's public QuestionPagedListPaginationQuery cursor, up to maxAnswersPerQuestion / maxAnswers / maxAnswerPages. Every paginated answer is checked against the page question (title/path, linkedQuestion, originalQuestionIfDifferent, and the answer permalink), so unrelated promoted answers cannot become dataset rows while legitimate merged variants still pass.
• Paginates comment areas and nested replies through AllCommentsListQuery (10+ pages with after cursors), with maxCommentsPerAnswer as the hard cap; this is opt-in and robots-sensitive.
• Optionally reads a Topic feed, a Space SSR multifeed, profile statistics, Topic statistics, and Post metrics.
• Optionally reads the public question /log audit for created time, view count, and asker identity (includeQuestionAudit), and paginates profile answer/question history (includeProfileHistory).
• Optionally resolves public identity lists for question followers, answer upvoters, profile followers/following, and Space roles through explicitly requested relationship types and the sensitive-data gate.
• Walks the question → answer → comment chain and keeps the links between them (entity_edge rows), so you can rebuild the conversation tree.
• Offers four output shapes via outputFormat: normalized (separate rows per entity, default), nested (answers grouped under their question, comments under their answer), flat, and edges (relationship rows plus any status rows).
• Reports a status row (rather than guessing) whenever a page is blocked, login-gated, removed, or otherwise inaccessible, and a run_manifest row summarizing the whole run.
• Leaves empty content fields out of each row rather than filling them with a guess. 0 and false are kept as real values.
• Never emits cookie/credential material. URLs are redacted when redactSensitiveQueryParams is on.

Output per record

Every row carries content_type (question, answer, comment, profile, profile_activity, topic, topic_post, space, space_feed_item, entity_edge, relationship_collection, status, or run_manifest), plus shared fields: record_id, title, url, canonical_url, source_url, access_state, scrape_timestamp, requested_depth, observed_depth, detail_expansion_state, output_format, matched_input, matched_input_kind, and (when search seeds were used) source_query/source_queries.

Every row also carries run-context fields describing how it was collected: source_observations/source_observation_count, redirect_chain, url_redaction_state/url_redactions, proxy_used, proxy_path_used, proxy_resolution_state, session_cookie_count, session_input_state, session_cookie_expiry_state, request_delay_seconds, browser_locale, requested_language, navigation_retries/navigation_retry_delays_seconds, parser_version and parser_strategy. With includeFieldEvidence on (default), rows also get field_sources (where each value came from) and fields_present.

access_state is public for normal pages. Other values include login_required, cloudflare_challenge, rate_limited, not_found, removed, unavailable, quora_plus_or_gated, quora_plus_preview, redirected_to_search, unsupported_url, search_unavailable, answer_pagination_limited, answer_pagination_unavailable, comments_unavailable, comments_capped, topic_feed_unavailable, topic_feed_capped, space_feed_capped, profile_history_limited, profile_history_unavailable, question_audit_unavailable, relationship_identity_unavailable, relationship_identity_capped, relationship_identities_not_observed, state_store_unavailable, state_store_uncommitted and network_error.

Question row — question_url, question_id, question_answer_count (label) / question_answer_count_value (parsed number) / question_answer_count_scraped (rows actually collected), question_follower_count / question_follower_count_value (when the public payload exposes it), related_questions (qid, title, url), question_created_at, question_updated_at, question_view_count, question_audit_state, question_merge_count, question_age_days/question_age_state, last_answered_at + last_answered_at_evidence, question_author_name/_url/_image_url/_id/_is_verified/_is_anonymous/_credentials (from a visible author card or the opt-in /log audit), question_image_urls/_video_urls/_outbound_links (opt-in), answer_count, answer_limit_reached, topics, topic_links.

Answer row — answer_id, answer_node_id, answer_url, answer_text, answer_html (opt-in, derived from rich text), answer_media_urls/_image_urls/_video_urls/_outbound_links (opt-in), answer_translation_state/_label, answer_original_url, answer_is_anonymous, answer_author_is_verified, answer_author_follower_count (when present), answer_created_at/_updated_at, answer_age_days/_state, answer_visibility, answer_text_completeness, answer_is_collapsed, answer_is_quora_plus, answer_access_option, answer_effective_access, answer_is_top_labeled, answer_rank/_rank_source, answer_upvote_count/_value, answer_comment_count/_value, answer_view_count/_value, answer_share_count/_value, reaction_breakdown (opt-in), is_ai_answer/ai_label_text/ai_label_source, related_content/_count/_observed_count/_limit_reached (opt-in), answer_jsonld_type, author_name/_url/_image_url/_id/_credentials.

Comment row — comment_id, comment_url, comment_text, comment_source (graphql_comment_area), comment_author_name/_image_url/_id/_credentials/_is_verified, comment_is_anonymous, comment_timestamp/_created_at, comment_age_days/_state, comment_visibility, comment_upvote_count/_value, comment_reaction_count/_value, comment_reply_count/_value, comment_reaction_breakdown, comment_media_urls/_image_urls/_video_urls/_outbound_links, comment_is_top_labeled, comment_rank/_rank_source, comment_sort_order, comment_depth, comment_depth_limit_reached, comment_is_reply, parent_comment_id/_url, comment_reply_path, comment_container_id, comment_items_observed_count/_emitted_count, comment_limit_reached, comment_page_limited.

Profile row — profile_url, profile_name, profile_bio, profile_image_url, profile_author_id, profile_is_verified, profile_is_restricted, profile_account_status, profile_is_anonymous, profile_is_deceased, profile_snapshot_state, plus profile_follower_count, profile_following_count, profile_answer_count, profile_question_count, profile_post_count, profile_total_content_views, profile_month_content_views, profile_joined_at, profile_credential, profile_work_credentials, profile_school_credentials, profile_location_credentials, profile_expertise_topics, profile_history_limited when the visible feed is truncated or the opt-in history is capped, and profile_history_state (not_requested/observed/capped/rate_limited/unavailable/empty) when history is requested.

Profile-activity row — activity_url, activity_type, activity_source, activity_question_url/_title/_id, activity_answer_id, created_at, author_name/_url/_image_url/_id/_credentials.

Topic row — topic_url, topic_id, topic_follower_count, topic_most_viewed_authors (uid, name, url, follower_count, num_views, num_public_most_viewed_answers, credential), topic_photo_url, topic_default_photo_url, topic_has_leaderboard, and topic_feed_state (not_requested/observed/capped/unavailable).

Topic-Post row — post_url, post_pid, post_node_id, post_title (explicit title or first content line), post_body, post_media_urls/_image_urls/_video_urls/_outbound_links (opt-in), post_author_name/_url/_image_url/_id/_credentials/_is_verified/_is_anonymous/_follower_count, post_question_url/_title, topic_url, post_created_at/_updated_at, post_age_days/_state, post_updated_age_days/_state, post_upvote_count, post_view_count, post_share_count, post_comment_count, post_effective_access, post_is_question_summary.

Space row — space_url, space_id, space_follower_count, space_all_time_followers (conditional), space_member_count, space_contributor_count, space_moderator_count, space_admin_count, space_description, space_icon_url, space_cover_url, space_item_count, space_feed_capped, space_ads_skipped.

Space feed-item row — feed_url, feed_type, feed_title, feed_text (opt-in), feed_post_pid/feed_post_node_id (post stories), feed_author_name/_url/_image_url/_id/_credentials/_is_verified/_is_anonymous, feed_created_at, feed_age_days/_state, feed_upvote_count, feed_view_count, feed_share_count, feed_comment_count, feed_effective_access, feed_is_ai_answer, feed_question_url/_title, media_urls/image_urls/video_urls/outbound_links (opt-in).

Relationship row (entity_edge) — relationship, edge_id, from_record_id/_content_type/_url/_id, to_record_id/_content_type/_url/_id/_name/_image_url, node_image_url, edge_state, unresolved_reason, edge_evidence_field, relationship_collection_key, collection_total_text/_value.

Relationship-count row (relationship_collection) — relationship_collection_key, collection_total_text, collection_total_value, edge_evidence_field.

Status row — access_state, http_status, error, source_url, and for pages that failed to load final_url, scrape_duration_ms, navigation_attempts, navigation_retries. Run-level status rows (url: run://quora-dataset-builder) report a state store that could not be read (state_store_unavailable), state that was not saved because a source did not finish (state_store_uncommitted), or a run that produced no real rows (no_records_emitted).

Option-dependent fields — with a stateStoreId, entity rows add event_type (new/changed/unchanged/scope_mismatch), state_snapshot_hash, previous_snapshot_hash, state_scope_fingerprint, state_store_state, only_new_items, include_unchanged. With includeRawEvidence, page-level rows add raw_html_sha256, raw_html_bytes, raw_html_truncated, raw_html_returned_bytes, raw_html_redaction_state, raw_html_redactions, raw_evidence_state, raw_evidence_retention_days, plus raw_html when includeHtml is also on and rawEvidenceRetentionDays is above zero. In nested format, question rows carry answers arrays (each answer with its comments) and nested_* count fields. With maxItemsPerSource set, rows add max_items_per_source and per-source cap fields.

Run-manifest row — one per run, summarizing counts and the settings actually applied: run_started_at/_finished_at/_duration_ms, records_emitted, source_pages_attempted, content_type_counts, access_state_counts, input_accounting, requested_result_types, requested_output_fields, dedupe_mode, duplicate_records_suppressed, proxy_path_used, proxy_resolution_state, search_discovery_source, search_discovery_hits, answer_gql_pages, answer_pagination_state, answer_pagination_unrelated_suppressed, comment_gql_pages, topic_feed_pages, space_feed_bundles, quora_plus_preview_count, comment_cap_hit_count, topic_feed_cap_hit_count, profile_history_pages, profile_history_state, relationship_identity_pages, relationship_identity_requests, relationship_identity_states, question_audit_state, question_audit_attempts, legacy_input_notes, hard_limits, gql_calls, query_hashes, and the per-run values of your own input settings echoed back for reconciliation.

Trimming rows, depth, sessions and caps

• Field allowlist. outputFields keeps only the fields you name, applied after extraction; an empty list returns the full row. Nested projection is supported with dotted paths such as answers.answer_text or answers.comments.comment_text in nested output. Child identity/URL fields stay in place, and content_type, identity, access/status, depth, timing and parser fields are always protected so a trimmed row can still be traced. output_projection on each row reads full or selected, and the run manifest records the requested fields. includeFieldEvidence toggles the field_sources/fields_present maps.
• Depth. scrapeDepth: "detail" (default) runs the capped traversal; overview emits page-level summaries without answer, comment, profile-activity or Space-feed expansion. Every row reports requested_depth, observed_depth and detail_expansion_state. bounded means a configured cap was reached, not that Quora's full history was collected, and blocked pages report observed_depth: "status".
• Caps. maxAnswersPerQuestion bounds each question, maxAnswers is a global answer cap, maxAnswerPages bounds GraphQL pages explicitly, maxItems caps entity/status rows before graph/manifest rows, maxItemsPerSource caps each source, and maxGraphEdges caps relationship rows. When maxItems/maxGraphEdges is 0 the corresponding cap is off. maxRelationshipItemsPerCollection bounds each labeled relationship collection, maxRelationshipPages bounds each public identity-list query, and maxProfileHistoryPages bounds each profile answer/question list. answer_pagination_limited is emitted only when a configured page cap or a GraphQL rate limit stopped pagination; an exhausted source is not reported as a cap.
• Comments. Set includeComments: true. When maxCommentsPerAnswer is 0, the documented fallback is 10 per answer. The actor calls AllCommentsListQuery with after cursors and a page size derived from your cap (up to 50), so a 24-comment answer is collected as 10 + 10 + 4 when the server pages at ten. includeCommentReplies (or includeReplyGraph) expands nested replies up to maxCommentDepth; comment_page_limited is set only when a cap is actually hit, and a comments_capped status row follows. If the paginated query fails, the actor falls back to the first-page loader rather than inventing rows. Comments use Quora's comment-area GraphQL, which robots.txt disallows for generic agents: keep this opt-in and review your own compliance posture.
• Topic feed. Set includeTopicFeed: true to call the topic page's own public GraphQL feed. maxTopicFeedPages bounds pages. This path is also robots-sensitive; keep it opt-in.
• Question audit. includeQuestionAudit: true reads the question's public /log page for question_created_at, question_view_count, question_merge_count, and the asker name/uid/profile/credential. It retries a transient 403 once, then keeps question_audit_state: "unavailable" and a question_audit_unavailable status row. /log is robots-disallowed for generic agents and is never part of the daily prefill.
• Profile history. includeProfileHistory: true paginates UserProfileAnswersMostRecent_RecentAnswers_Query and UserProfileQuestionsList_Questions_Query and emits bounded profile_activity rows. maxProfileHistoryPages (default 0 = three pages per list) and requestDelay (used as a 1.5–2.5 s GraphQL pace when above zero) bound the work. A cap sets profile_history_limited; a 400/403 sets profile_history_state: "rate_limited" and a profile_history_unavailable status row instead of fabricated rows.
• Identity lists. Identity rows are emitted only when relationshipTypes explicitly requests them, the matching family flag is on, and includeSensitiveRelationshipData is on for sensitive types. The public GraphQL operations cover profile followers/following people/questions/Spaces/Topics, question followers, answer upvoters, and Space admins/contributors/moderators/followers. maxRelationshipItemsPerCollection and maxRelationshipPages bound every list; Space role calls run from a www.quora.com page context. These lists are robots-sensitive and rate-sensitive; a failure keeps a relationship_identity_unavailable or relationship_identities_not_observed status row, and a page/item cap adds relationship_identity_capped.
• Sessions and proxies. No account, login, cookie, or session secret is used. The transport ladder is direct (curl_cffi, fresh session per attempt) → optional user proxy override → Apify AUTO datacenter → RESIDENTIAL only as a hardcoded last resort after the free rungs fail. A missing proxyConfiguration never stops the run, and there is no paid default. proxy_path_used reports the rung that worked.
• Exact field names. Every output field this actor can emit is listed under Output per record above. Fields whose source is absent on a given page are omitted, not set to null or a sentinel.

Complete schema input inventory

Every input is optional. The daily prefill uses one public question URL with a small answer cap, comments and feeds off, and no paid proxy.

Field	Type	Default	Description
mode	string	auto	auto / byQuestionUrls / byAnswerUrls / byProfileUrls / byTopicUrls / bySpaceUrls / byPostUrls / search; typed modes validate seed routes
scrapeDepth	string	detail	overview (page-level summary only) or detail (full capped traversal)
startUrls	array	[]	Question, answer, profile, Topic, Space, or Post URLs; mixed lists are auto-detected and invalid hosts become status rows
searchQueries	array	[]	Best-effort discovery terms; Quora's native search is login/Cloudflare-gated and is never claimed
questionUrls	array	[]	Typed question seeds
answerUrls	array	[]	Typed answer-permalink seeds
topicUrls	array	[]	Typed Topic seeds
spaceUrls	array	[]	Typed Space seeds; automatically enables Space output
profileUrls	array	[]	Typed profile seeds
postUrls	array	[]	Typed standalone Post seeds
resultTypes	array	["question","answer","comment","profile","profile_activity","topic","topic_post","space","space_feed_item","entity_edge","status","run_manifest"]	Which content_type values to include
outputFormat	string	normalized	normalized / nested / flat / edges
dedupeMode	string	canonical	canonical (merge duplicate entities by canonical URL/ID) or none
onlyNewItems	boolean	false	With stateStoreId, suppress rows unchanged since the last run
includeUnchanged	boolean	true	Keep unchanged rows when onlyNewItems is on
stateStoreId	string	-	Apify key-value store name for incremental re-scraping; required when onlyNewItems is on
outputFields	array	[]	Optional field allowlist for smaller rows; identity/status fields are always kept
includeFieldEvidence	boolean	true	Include field_sources/fields_present on each row
includeUnresolvedEdges	boolean	true	Keep relationship rows whose target could not be matched
relationshipTypes	array	[]	Allowlist of labeled relationship collections to expand (follower, following, want-answers, upvoter, Space role); identity-revealing types also need includeSensitiveRelationshipData and the matching family opt-in
includeProfileEdges	boolean	false	Per-family opt-in for profile follower/following/question/space/topic relationships
includeQuestionFollowerEdges	boolean	false	Per-family opt-in for question follower/want-answer identities (sensitive)
includeAnswerUpvoterEdges	boolean	false	Per-family opt-in for answer upvoter identities (requires includeSensitiveRelationshipData)
includeSensitiveRelationshipData	boolean	false	Privacy gate required for identity-revealing relationship collections
includeTargetMetadata	boolean	false	Keep the visible name/image of each relationship target
includeRelationshipCounts	boolean	false	Emit privacy-preserving member-count rows; also add relationship_collection to resultTypes
includeSpaceRoleEdges	boolean	false	Keep Space contributor/moderator/admin edges
includeSpaceContentEdges	boolean	true	Keep Space to post/question membership edges and feed-item membership
maxRelationshipItemsPerCollection	integer	500	Cap per labeled relationship collection
maxQuestions	integer	25	How many discovered question links get queued on top of your own seed URLs
maxAnswersPerQuestion	integer	10	Maximum answer rows kept per question after SSR and GraphQL pagination
maxAnswers	integer	0	Global answer cap across the whole run (0 = no cap)
includeQuestion	boolean	true	Keep question entities in the output
includeAuthor	boolean	true	Keep author/profile entities and author fields in the output
includeAuthorSnapshot	boolean	false	Fetch a bounded profile snapshot for each visible answer author
includeTopics	boolean	true	Keep topic entities and topic links
includeSpaces	boolean	false	Keep Space and Space feed-item rows; switched on automatically when a Space URL is used in startUrls or spaceUrls
includeComments	boolean	false	Opt-in comment rows with AllCommentsListQuery cursor pagination; robots-sensitive
includeEngagementBreakdown	boolean	true	Include per-reaction breakdowns alongside aggregate counts
includeRelatedContent	boolean	false	Collect explicitly attached answer related-content links when Quora's public payload exposes them; otherwise they are omitted
maxRelatedContent	integer	20	Bound related-content links per answer when enabled; zero emits no related-content items
topCommentsOnly	boolean	false	Keep only comments explicitly labeled Top/Most helpful
commentSortOrder	string	observed	observed / upvotes_desc / reactions_desc / replies_desc / newest
includeReplyGraph	boolean	false	Compatibility alias that enables comments and nested replies
includeMedia	boolean	false	Include image/video URL fields where available
includeFeedText	boolean	false	Include visible text on Space feed cards
includeOutboundLinks	boolean	false	Include non-Quora links found in rich text
includeEntityGraph	boolean	false	Emit relationship rows even when entity_edge is not in resultTypes
maxCommentsPerAnswer	integer	0	Comments kept per answer; when comments are on and this is 0, the fallback is 10; the cap is the pagination limit, not a page size
maxCommentDepth	integer	20	Maximum nested reply depth when replies are enabled; 0 disables replies
maxItems	integer	0	Cap on entity/status rows before graph and manifest rows (0 = no cap)
maxGraphEdges	integer	5000	Cap on relationship rows (0 = no edges)
maxScrollAttempts	integer	5	Scroll attempts per page in detail mode
maxLoadMoreAttempts	integer	0	"Load more" button clicks per page
includeRawEvidence	boolean	false	Attach a document hash and size/redaction metadata for auditing
includeHtml	boolean	false	Retain bounded raw HTML on page-level rows and derived answer HTML
maxHtmlBytes	integer	100000	Upper bound for retained or derived HTML
requestDelay	number	1	Delay in seconds between page batches
maxConcurrency	integer	1	Independent seeds processed in parallel (values above 3 are treated as 3)
maxNavigationRetries	integer	2	Retries after a failed page load (max 3)
language	string	auto	Browser language hint (en, es, fr, ...); not a translation guarantee
proxyConfiguration	object	-	Optional proxy override. The free ladder runs first; there is no paid default
redactSensitiveQueryParams	boolean	true	Redact credential-like query values and fragment credentials from emitted URLs/HTML
rawEvidenceRetentionDays	integer	30	Retention label for raw HTML evidence; 0 keeps metadata only
maxRawEvidenceBytes	integer	100000	Size bound for raw HTML evidence
maxItemsPerSource	integer	0	Cap on records from each individual source URL/query (0 = no cap)
searchDiscovery	string	relatedOnly	relatedOnly (related questions only) / bestEffortIndex (third-party index) / off
maxRelatedQuestions	integer	20	Related questions enqueued per question seed
searchResultsPerQuery	integer	10	Maximum best-effort index results per query
maxAnswerPages	integer	0	Explicit answer GraphQL page cap; 0 derives from maxAnswersPerQuestion
maxTopicFeedPages	integer	3	Topic-feed GraphQL page cap when includeTopicFeed is on (minimum one page)
maxSpaceFeedBundles	integer	4	Space feed bundles read from the SSR multifeed
includeCommentReplies	boolean	false	Expand nested repliesConnection comment replies
includeQuoraPlusPreview	boolean	true	Emit Quora+ preview answers with preview flags; when false, omit them
includeProfileStats	boolean	true	Emit profile follower/following/answer/question/post/view/joined/credential fields
includeTopicStats	boolean	true	Emit Topic follower count and most-viewed authors
includeSpaceStats	boolean	true	Emit Space counters, description, icon and cover
includePostStats	boolean	true	Emit Post counters, effective access and question-summary flag
includeAnswerStats	boolean	true	Emit answer counters, dates, AI flag and effective access
includeTopicFeed	boolean	false	Browser GraphQL Topic feed; opt-in and robots-sensitive
includeQuestionAudit	boolean	false	Read the public question /log for created time, view count, merge count, and asker identity; robots-sensitive opt-in
includeProfileHistory	boolean	false	Paginate the public profile answer/question tabs into bounded profile_activity rows; robots- and rate-sensitive opt-in
maxProfileHistoryPages	integer	0	Cap per profile history list; 0 derives three pages (about 30 items per list)
maxRelationshipPages	integer	3	Cap per public identity-list query after the relationship allowlist and privacy gates

Example inputs

Example: full detail for a question

{
  "questionUrls": [{ "url": "https://www.quora.com/What-is-the-best-way-to-learn-Python" }],
  "maxAnswersPerQuestion": 20,
  "includeComments": true,
  "maxCommentsPerAnswer": 10
}

Example: best-effort search discovery, compact nested output

{
  "searchQueries": ["machine learning career advice"],
  "searchDiscovery": "relatedOnly",
  "maxQuestions": 10,
  "outputFormat": "nested",
  "maxAnswersPerQuestion": 5
}

Example: a Space's content and follower count

{
  "spaceUrls": [{ "url": "https://www.quora.com/q/example-space" }],
  "resultTypes": ["space", "space_feed_item", "relationship_collection", "run_manifest"],
  "includeRelationshipCounts": true,
  "relationshipTypes": ["space_has_contributor"]
}

Example: incremental re-scrape of the same question

{
  "questionUrls": [{ "url": "https://www.quora.com/What-is-the-best-way-to-learn-Python" }],
  "onlyNewItems": true,
  "includeUnchanged": false,
  "stateStoreId": "python-question-watch"
}

Use cases

• Content research — pull full question-answer-comment threads on a topic for analysis
• Community monitoring — re-scrape the same questions on a schedule and keep only rows flagged new/changed
• Knowledge graph building — use outputFormat: "edges" to load the question/answer/topic relationship graph into a graph database
• Influencer and expert tracking — follow a profile's visible answers and its public follower/following/content statistics
• Space analysis — collect a Quora Space's feed, metrics and contributor counts

FAQ

Do I need a Quora account? No. Direct public question, answer, profile, Topic, Space, and Post pages are read without login. The actor never asks for, stores, or emits cookies or session secrets.

Why did I get a status row instead of question/answer data? Quora showed a login wall, Cloudflare challenge, a removed page, or an unsupported redirect for that URL. The actor reports this honestly as a status row instead of guessing at data that was not there. A run with zero real rows also gets a no_records_emitted status row and a non-success status message.

Why do I sometimes get fewer answers than maxAnswersPerQuestion? The actor reads the server-rendered answers, filters out answers whose question is provably unrelated, then paginates the public GraphQL cursor up to your cap. If the question genuinely has fewer answers, the cursor stops; you get the answers that were observable and no answer_pagination_limited row, because the source was exhausted rather than capped. answer_pagination_limited appears only when maxAnswerPages/your cap stopped a non-exhausted source or Quora rate-limited a pagination request, and the page's known answerCount is not already satisfied; a non-rate GraphQL failure uses answer_pagination_unavailable. Neither case fabricates an answer.

How do I get question created time, view count, or the asker? Set includeQuestionAudit: true. The actor reads the public /<question>/log page (robots-sensitive, opt-in) and fills question_created_at, question_view_count, question_merge_count, and the question_author_* fields. Without it, those fields are omitted rather than guessed. A blocked /log leaves question_audit_state: "unavailable" and a status row.

How do I get more than the three visible profile history items? Set includeProfileHistory: true. The actor paginates the public answer/question tab connections into profile_activity rows, bounded by maxProfileHistoryPages and paced by requestDelay. This GraphQL path is robots- and rate-sensitive; a 400/403 leaves a profile_history_unavailable status row and no fabricated rows.

What is the difference between the four outputFormat values? normalized keeps every entity as its own row (the default). nested groups each question's answers (and each answer's comments) underneath it; rows without a matching parent stay at the top level with an orphan reason. flat is like normalized, except answer rows also get question context and comment rows get answer_url. edges returns the relationship rows between entities plus any status rows.

Can I get a Quora account's followers, question followers, answer upvoters, or Space roles? Yes, as explicit opt-ins. Request the relationship type in relationshipTypes, turn on the matching family flag (includeProfileEdges, includeQuestionFollowerEdges, includeAnswerUpvoterEdges, or includeSpaceRoleEdges), and set includeSensitiveRelationshipData: true for follower/upvoter/want-answer types. The actor then calls the public identity-list GraphQL operations, paginates up to maxRelationshipPages, and emits only identities whose profile/question/Space/Topic URLs it observed. It may also use visible DOM links as a fallback. These paths are robots- and rate-sensitive; failures produce relationship_identity_unavailable or relationship_identities_not_observed status rows. It never invents identities.

What does the incremental re-scrape (onlyNewItems) actually compare? It hashes each entity's semantic content and compares it to the previous run stored under stateStoreId, tagging each row new, changed, unchanged, or scope_mismatch (the previous run used different settings). Rows are only dropped when onlyNewItems is on and includeUnchanged is off. Status and manifest rows are never dropped by this setting. If any requested source did not finish, the new state is not saved and a state_store_uncommitted status row tells you so.

Why are some text/media fields missing even with includeMedia/includeHtml on? Those fields are only populated when Quora's public payload actually contains that content for the entity in question. The toggles request the data; they do not guarantee every entity has it.

Is search native Quora search? No. Quora's native /search?q= is login/Cloudflare-gated for logged-out clients and is disallowed by robots.txt. searchQueries uses related-question discovery by default and an optional third-party best-effort index; if no public results are discoverable it emits a search_unavailable status row.

What are the hard limitations? Native Quora search is not available cookielessly: /search?q= redirects and the public search GraphQL returns no connection. Quora+ smart_distribution answers may be preview/truncated; they are flagged answer_is_quora_plus and their observed completeness is reported, and no gated remainder is fabricated. Space all-time views/upvotes/comments are private (tested across ten diverse Spaces), Topic description and question count are absent, and question description is absent. Space allTimeFollowers is emitted as space_all_time_followers only when the payload carries it (not every Space does).

Question created/view/asker, full profile history, comment/reply pagination, and identity lists are technically available through public pages/GraphQL but are robots-sensitive and/or rate-sensitive opt-ins; they are off by default and not part of the daily prefill. Be aware that robots.txt for generic agents disallows / broadly, including /*_POST, /*/log, /*/followers, /*/following, and /*/comments; product/legal review decides whether to enable those opt-ins. Quora URLs are valid in a browser, but plain HTTP clients can receive a Cloudflare 403 on HEAD/GET; the actor fetches them with TLS impersonation.

Is this affiliated with Quora? No. This is an independent, third-party actor that reads public Quora pages; it is not affiliated with or endorsed by Quora, Inc. You are responsible for how you use the data.