Promptwatch

Create or reuse a brand, link it to the project, and return the project-brand row. Setting relation to SELF demotes the previous SELF brand.

createBrand

Queue quota-gated, asynchronous content generation. `CREATE` makes new content; `OPTIMIZE` creates a rewrite from stored crawl content without editing or publishing the live page. Returns `{ id, status }`; poll `getContent` until COMPLETED, FAILED, or STOPPED.

createContent

Create an active monitor and return its record with promptCount. Use it before adding prompts; no prompts are created, and persona data is snapshotted.

createMonitor

Create an audience persona and return its row. Use it to shape future monitors; names are unique per project.

createPersona

Create a Promptwatch project and return its record, including id and slug. Requires an organization-level writable credential and project quota; website enrichment runs asynchronously.

createProject

Create one tracked prompt and return its row. Enforces active-prompt quota and queues response processing asynchronously.

createPrompt

Create prompts for one monitor and return their rows. Enforces active-prompt quota and queues response processing asynchronously.

createPromptsBulk

Queue an asynchronous PDF Report for the project. Supports dashboard and monitor types with the same filters as the Reports UI (email is never sent). Returns `{ reportId, status }`; poll `getReport` until complete for a Download URL.

createReport

Create missing tags and return matching tag rows. Names are trimmed and deduplicated; this does not attach them.

createTags

Create missing topics and return matching rows. Names are trimmed, lowercased, and deduplicated; this does not attach them.

createTopics

Deactivate prompts and return { count }. Stops future scheduling while retaining response history.

deactivatePrompts

Decline a REVIEW, APPROVED, or SCHEDULED piece. Requires dashboard setup. The slot becomes SUPPRESSED and will not be generated or published.

declineContentSlot

Soft-delete and deactivate a monitor, returning { success: true }. Stops future scheduling but retains historical data.

deleteMonitor

Delete a persona and return { success: true }. Existing monitors retain their snapshotted values.

deletePersona

Soft-delete and deactivate a prompt, returning { success: true }. Future runs stop; response history is retained.

deletePrompt

Soft-delete and deactivate prompts, returning { count }. All ids must belong to the project; response history is retained.

deletePromptsBulk

Stop tracking a product; extracted shopping data is unaffected. An unknown product returns SHOPPING_TRACKED_PRODUCT_NOT_FOUND. Requires the Shopping insights plan feature.

deleteShoppingTrackedProduct

Delete a tag and return { success: true }. Prompt links cascade; other references may block deletion.

deleteTag

Delete a topic and return { success: true }; its prompt links are removed.

deleteTopic

Summarize the project's action board: counts by status/severity/type, a catalog of every action type with open counts (suggested/todo/in_progress only — done is excluded), recent non-dismissed actions, and stale suggested/todo items. Call this before creating actions to avoid duplicates.

getActionBoardOverview

Get the top advertiser domains with daily ad counts (defaults to the last 90 days when dates are omitted). Returns `{topDomains,daily}` for share-of-ads trends. Requires the Ads Radar plan feature.

getAdDomainAnalytics

Return { timeSeries, topBrands } for brand visibility comparisons. Supports model, prompt, monitor, type, relation, and name/domain filters; omitted dates query today only, inclusive UTC.

getBrandVisibilityOverTime

Return `{dates,series,totalCitations}` for citation counts and share by domain over time. Use to compare sources; `averageRank` is within-answer position where 1 is best.

getCitationDomainsOverTime

Return each model/provider with `domains:[{domainName,count}]` citation totals. Use to compare source preferences across AI engines; `limit` applies per model.

getCitationDomainsByLlm

Return `{llm,provider,model,totalCitations,uniqueDomains,uniquePrompts}` rows. Use to compare citation breadth by AI engine; `limit` caps model rows.

getCitationLlmSources

Return `{timeSeries,overallAverageRank,lastRank}` for within-answer positions of self-domain citations. Use to track citation prominence; rank 1 is best, not most-cited-domain rank.

getCitationRankAnalysis

Return cited-domain/URL frequencies plus total estimated monthly traffic and average authority/PageRank metrics. Use for a source overview; `domains.data` is limited, while `urls` is not.

getCitations

Return `{brandId,competitor,model,percentage}` cells comparing brand visibility by model. Use for cross-model benchmarking; percentage is a 0–100 visibility score, not sentiment.

getCompetitorHeatmap

Get one generated document or poll a `createContent` job. Returns status, configuration, prompt, markdown `content`, failure details, and timestamps; content may be null until completion.

getContent

Get the project's Content Agent configuration (autonomy, publish windows, budget share, CMS). Returns `settings: null` until setup has been completed once.

getContentAgentSettings

Get one scheduled Content Agent piece including draft markdown, document configuration, and decline reason when present.

getContentSlot

Get a prompt's latest stored analysis and recommendations for creating or optimizing content. Returns prompt context, current 0-100 coverage, report/source data, and recommendation IDs/details; null means no analysis.

getContentGapRecommendations

Summarize current content coverage across active prompts. Returns `averageCoverageScore`, analyzed/unanalyzed totals, distribution, and type/intent breakdowns; scores use each prompt's latest stored 0-100 analysis, and missing analysis is not zero.

getContentGapStats

Get a document's CMS publish history, newest first, plus `latest` as a shortcut to the most recent record. `latest` is null when the document has never been pushed. Use it to check whether a draft is STAGED before publishing live.

getContentPublishStatus

Shows AI crawler requests over time from ingested server/CDN logs. Returns isSuccess, isEmpty, data, and categories; these are machine requests, not visits or citations. When isEmpty is true, no crawler logs have been recorded and data and categories are empty.

getCrawlerTrend

Return paginated `{data,total,page,size}` rows grouped by cited URL, with counts, 1-based positions, model coverage, and SEO metrics. Use for citation audits; position 1 is best and SEO ranks are separate metrics.

getGroupedCitations

Get a prompt's latest stored content-gap analysis. Returns `contentCoverageScore`, report, sources, sitemap counts, and recommendations; the score is 0-100, while null means no analysis.

getContentGapLatest

Return `{date,brandMentions,competitorMentions}` buckets for collected responses. Use to compare mention trends; competitorMentions counts occurrences, not unique brands.

getMentionsTimeSeries

Get one monitor with configuration and metrics, including promptCount, responseCount, averageVisibility, and responseCountInRange. Omitted dates query today only, inclusive UTC.

getMonitor

Gets one tracked URL with AI-response citation totals for a period. Returns page details; dates default to today-only UTC and reject future days, while zero citations in-range can return TRACKED_PAGE_NOT_FOUND.

getPageTracker