assign_to_campaign
ChatGPTAdd up to 500 existing posts to a campaign. Grouping only — nothing about a post’s content or schedule changes. A write action — confirm with the user before calling. Requires the campaigns.manage scope.
bulk_post_target_action
ChatGPTReschedule, retry, cancel or delete up to 5000 post TARGETS in one call — the batch form of reschedule_post / retry_post / cancel_post; use ONE call for the same change on many posts, never one call per post. action is reschedule | retry | cancel | delete. For reschedule, pass either params.scheduledAt (an absolute instant) or params.shiftMinutes (relative), optionally with params.respectCadence; a relative shift is all-or-nothing across the batch. cancel stops a post and keeps its row, which stays in the list as Canceled — it is how a failed post is cleared. delete removes a row, but only one that was never attempted: published posts, posts held on the platform’s own scheduler and attempted posts (most failed ones) are skipped and named in skippedReasons, never rewritten to cancel. Returns { jobId, accepted, skipped, skippedReasons? } — a skipped target is reported, never silently dropped, and accepted: 0 means nothing changed. This changes real scheduled content — confirm with the user before calling. Requires the posts.compose scope.
cancel_post
ChatGPTCancel a post target that has not gone out (draft, scheduled, on the platform’s own timer, on hold, or failed); a published, canceled or missed target cannot be canceled. This stops a real scheduled post — confirm with the user before calling. Requires the posts.compose scope.
clear_post_recycle
ChatGPTTurn OFF evergreen re-posting for a post. Idempotent — clearing a post that was not enrolled succeeds. Already-scheduled recycled copies are not retracted by this; cancel those with cancel_post or bulk_post_target_action. A write action — confirm with the user before calling. Requires the posts.compose scope.
create_campaign
ChatGPTCreate a campaign — a named group of posts with an optional client, colour and date range. Posts are added with assign_to_campaign. A write action — confirm with the user before calling. Requires the campaigns.manage scope.
create_client
ChatGPTCreate a new client group (a brand or influencer whose accounts are managed together). This adds a real client to the workspace — confirm with the user before calling. Requires the clients.manage scope.
create_connect_invite
ChatGPTMint a shareable self-connect link so a client can connect the requested platforms themselves. Returns the link + expiry; share it with the client. This creates a real invite — confirm with the user before calling. Requires the accounts.connect scope.
create_pinterest_board
ChatGPTCreate a board on a connected Pinterest account and return its { id, name }, ready to use as the boardId of a Pin. privacy is PUBLIC (default) or SECRET. A write action — confirm with the user before calling. Requires the posts.compose scope.
create_pinterest_board_section
ChatGPTCreate a section on one board of a connected Pinterest account and return its { id, name }. A write action — confirm with the user before calling. Requires the posts.compose scope.
create_post
ChatGPTCompose a post and fan it out across the selected accounts (or save a draft with saveDraft:true). This creates real outbound content — confirm with the user before calling. Retry-safe: retrying with IDENTICAL arguments replays the original result instead of creating a duplicate post (the idempotency key is derived from the arguments). A Pinterest variant needs variants[].overrides.boardId: the board numeric id (digits only; list_pinterest_boards returns each board as { id, name }), or, for a post to exactly ONE Pinterest account, the exact board name, which is turned into the id. Any other board value is refused with pinterest_board_unresolved (422) before anything is saved: a name on a post to 2+ Pinterest accounts, a name no board has, or a name several boards have. Returns postId, status, targetCount, targets (one row per destination: targetId, platform, status, scheduledAt), replayed and, when there is something to say, warnings. A YouTube variant's variants[].overrides are typed and checked: an unknown key (override_key_unknown), a key another platform reads (override_key_wrong_platform), a deprecated spelling such as the single caption object, replaced by captions and accepted until 2027-04-30 (override_key_deprecated) and a wrong type on an older key (override_value_tolerated) are accepted and named in warnings, so read them; a wrong value on a newer key (captions, chapters, localizations, defaultLanguage, recordingDate) is refused with the field path (validation_failed). Requires the posts.compose scope.
create_posts_bulk
ChatGPTCreate up to 50 posts in one call. Each item is exactly the arguments of create_post, so a bulk item carries every override a single post does (per-platform caption, title, cover, per-asset alt text, first comment, per-platform settings), and a created item carries its own warnings. Items are created one after another, each on its own: one failing never rolls back the others, and the result lists every item in order with its post id or its error. Retry-safe: retrying with identical arguments creates nothing twice. Pass validateOnly:true to check every item without writing anything. It counts one request-budget unit per item. This creates real outbound content — confirm with the user before calling. Requires the posts.compose scope.
create_webhook
ChatGPTRegister an outbound webhook endpoint to receive HMAC-signed event deliveries (post.scheduled, post.published, post.failed, post.missed, target.on_hold, target.missed, account.disconnected, account.token_expiring, approval.requested, approval.decided, media.ready, report.completed). The signing secret is returned ONCE. This creates a real subscription — confirm with the user before calling. Requires the webhooks.manage scope.
decide_approval
ChatGPTApprove, reject, or request changes on a PENDING approval request, with an optional reviewer comment. Approving releases the gated post for publishing; an already-decided request returns a conflict. This gates real outbound content — confirm with the user before calling. Requires the approvals.decide scope.
delete_campaign
ChatGPTDelete a campaign. The POSTS in it are not deleted — they lose their campaign grouping and keep their own schedules. A write action — confirm with the user before calling. Requires the campaigns.manage scope.
delete_client
ChatGPTArchive a client group. OWNER ONLY — a Manager holding the clients.manage scope cannot do this. A write action — confirm with the user before calling.
delete_webhook
ChatGPTDelete an outbound webhook endpoint by id (its delivery history cascades). Idempotent. This removes a real subscription — confirm with the user before calling. Requires the webhooks.delete scope.
download_report_export
ChatGPTGet a time-limited download URL for one completed report run’s export file. The URL expires, so fetch it when you are ready to use it rather than storing it. Requires the reports.read scope.
edit_published_post
ChatGPTChange a video that is already live (or scheduled) on YouTube without republishing it: title, description, tags, category, privacy, kids / embedding / statistics settings, default language and translations, thumbnail and subtitle tracks. Send only the fields you are changing; an omitted field is left as YouTube holds it, and at least one is required. Takes the post id and the destination's targetId (from get_post or list_post_targets): the destination must be on YouTube and published or native_scheduled, otherwise 422 edit_not_supported. A natively scheduled video's privacy is set by its schedule: 422 edit_privacy_scheduled. A workspace under an operator's posting restriction gets 403 posting_restricted, because the edit writes to YouTube at once. Subtitle tracks are SRT or WebVTT with timing and replace the track of the same language and name, else are added. It spends YouTube API units (reading the video 1, the update 50, the thumbnail 50, a subtitle track 400, a replacement 450, plus 50 once to list the video's tracks) and is refused whole, before anything is sent, with 429 youtube_units_exhausted and a Retry-After in seconds to midnight Pacific time when the day's allowance is used up. The result lists what YouTube accepted in applied; a thumbnail or subtitle track YouTube refuses does not fail the edit and is explained in coverIssue, captions and captionIssue. A write action — confirm with the user before calling. Requires the posts.compose scope.
get_account
ChatGPTRead one connected social account in full — platform, handle, display name, avatar, status (including whether it needs re-authorisation, and reconnectReason — why it does: token_refresh_rejected, no_refresh_token, app_removed, missing_permission, security_change, credential_unreadable, repeated_auth_failures, moved_to_new_app, reconnect_requested or unknown), owning client and connection metadata, missingPermissions when a reconnect would add permissions the platform has since approved, and removedAt (an ISO timestamp) when the customer removed the channel from the console (it is disconnected and its history is kept). Requires the accounts.read scope.
get_account_analytics
ChatGPTAnalytics for ONE connected channel over an optional date window — the account grain that sits between get_analytics_overview (whole workspace) and get_post_analytics (one post). Returns totals (with totals.people: the channel’s separate people reached and accounts engaged for exactly the period, or null with the reason) and a daily series, plus two things available nowhere else: audience (profile-level activity the platform reports about the PROFILE rather than any post — profileViews, profileLinkTaps, videoViews, videoWatchTimeMinutes, contentPlays), and demographics (follower demographics as dimension → label → count, e.g. { country: { US: 120 }, age: { '25-34': 90 } }). demographics is a snapshot of the audience TODAY, not a window aggregate, and Meta suppresses it entirely below 100 followers — absent means not reported, never zero. Use list_accounts first to get a socialAccountId.
get_agency
ChatGPTGet the current workspace (agency) profile — name, plan, app mode, settings, and the caller’s role.
get_analytics_overview
ChatGPTGet the analytics overview over an optional date window and client / account filter. Returns summary (postsPublished, failed, onHold (destinations on hold, never counted as failed), impressions, facebookPosts (Facebook 3-second video views, Reel plays and views by format), clicks, likes, comments, shares, saves, engagementRate as a 0-1 fraction, costUsd, followers, followersChange (net change per account then added: value, base, accounts, and gained/lost for YouTube), followersGained, followersLost; for YouTube also watchTimeMinutes, avgViewDurationSeconds, subscriberFlow, trafficSources and viewerBreakdown, added over every video in the window), summary.people (reach and accountsEngaged: separate people, as the platform’s own figure for exactly the period and ONE channel, else value: null with the reason — people are never added across days, posts or channels), summary.audience when the platform reports profile-level activity (profileViews, profileLinkTaps, videoViews, videoWatchTimeMinutes, contentPlays), platforms[] (the same totals and people per platform, plus a daily impressions sparkline), accounts[] (per connected channel, with profileStats), and series[] (one entry per day). A metric a platform does not expose is omitted or null, never a fabricated 0. Separate people (reach, accounts engaged, engaged followers) are only in people: nothing in this answer adds them across posts, days or channels. Use get_post_analytics for per-post grain.
get_approval
ChatGPTRead one approval request with the gated post’s preview and roll-up, so a reviewer can see WHAT they are deciding on before calling decide_approval. Requires the approvals.read scope.
get_best_times
ChatGPTGet the worker-derived best posting times (strongest-first slots), optionally filtered by client / platform.
get_campaign
ChatGPTRead one campaign with its detail view — name, description, status, colour, date range, owning client, and its post counts. Requires the campaigns.read scope.
get_capabilities
ChatGPTThe per-platform composing rulebook: caption and hashtag limits, whether text-only is allowed, media rules per kind (allowed, whether OUR adapter can publish it, min/max count, accepted formats, file-size ceiling, duration bounds, aspect-ratio range), post types, native-scheduling window, default cadence, first-comment support, and the variants[].overrides keys each platform REQUIRES or accepts. Read this before composing instead of discovering a limit by failing a publish. Omit platform for all of them.
get_client
ChatGPTRead one client group in full — its name, type, approval policy, posting rules, brand voice and default timezone. Requires the clients.read scope.
get_client_analytics
ChatGPTOne client group’s analytics — totals, a per-platform breakdown and a time series. A different projection from get_analytics_overview filtered by client: this is the client rollup the console renders. Requires the analytics.read scope.
get_media_asset
ChatGPTRead one content-library asset in full — filename, kind, size, dimensions, duration, folder, labels, alt text, pipeline status, the publishable / ready flags plus the per-platform readiness map, a short-lived WebP thumbnailUrl and its placeholderColor (the preview dominant colour as #rrggbb, or null). Requires the media.read scope.
get_pinterest_top_pins
ChatGPTA Pinterest account’s best Pins over the last 7, 30 or 90 days, ranked by Pinterest itself and read live, so Pins made outside SkedCast are included. Each Pin has its link, title, a thumbnail, IMPRESSION / SAVE / PIN_CLICK / OUTBOUND_CLICK / ENGAGEMENT counts (null when Pinterest does not return one) and, when SkedCast published it, its postId. Use list_accounts to find the Pinterest account. Requires the analytics.read scope and the deepAnalytics plan feature.
get_post
ChatGPTGet a single post with its per-platform variants and resolved targets. Each target shows its status, holdReason, outcomeReason, failureReason (why a failed target failed: rejected, content_rule, retries_exhausted, internal_error or not_delivered) and requiredAction (reconnect_channel, change_plan, contact_support, reschedule, retry or check_platform; null when nothing is needed). A failed target whose outcomeReason is delivery_unconfirmed may already be live: its requiredAction is check_platform and retrying it is refused. A published target also carries the outcome of the steps after the publish as one sentence or null each — coverIssue, firstCommentIssue, captionIssue (the YouTube subtitle languages that were not added, and why) and playlistIssue — and firstCommentUrl, a link that opens a posted YouTube first comment (YouTube does not let an app pin it).
get_post_analytics
ChatGPTPer-post performance — complements get_analytics_overview with post grain. With postId: that post’s lifetime metrics broken down by target. Without: one keyset page of per-post rows (newest-first) over an optional date window and client / account / platform filter; pass the returned nextCursor to page. Every row carries impressions, reach, likes, comments, shares, saves, clicks, engagementRate and permalink. Where the platform reports them it also carries reactionsByType (Facebook: like/love/care/haha/wow/sad/angry/thankful/pride/fire/hundred), clicksByType, profileActivity, navigation (Instagram Stories: forward/back/exit), follows, profileVisits, reposts, quotes, linkClicks, totalInteractions, videoViews, videoCompletions, skipRate, watchTimeMinutes, avgViewDurationSeconds, avgViewPercentage, subscribersGained/Lost, audienceRetention, trafficSources and audienceBreakdown. null means the platform does not report that metric for this post — it never means zero. platformRemovedAt is set when the post was deleted on the platform; its figures are then final. In postId mode, totals and history add the post’s targets except reach, which counts people (a person reached on two channels is one person): read each target’s byTarget[].totals.reach.
get_post_recycle
ChatGPTRead one post’s evergreen re-posting configuration — interval, times, end date, remaining recycles — plus the worker’s bookkeeping (when it last recycled, how many times). Returns null when the post is not enrolled. Requires the posts.read scope.
get_post_target
ChatGPTRead a single fan-out destination in full — its status, scheduled time, platform, account, published permalink and external id, its failureReason when it failed (rejected, content_rule, retries_exhausted, internal_error or not_delivered; null otherwise, and null when the result of the send is unknown), holdReason when it is on hold (why: the platform’s posting limit, the channel’s gap, a slowdown to protect the channel, a plan budget, media still preparing, the plan not covering the channel, a channel to reconnect …; null otherwise), outcomeReason when it was missed (catch_up_window, plan_limit, channel_needs_reconnect) or when it failed and the post may already be live (delivery_unconfirmed: check the account first, retry is refused; null otherwise), requiredAction (what you must do: reconnect_channel, change_plan, contact_support, reschedule, retry or check_platform; null when nothing is needed), its thumbnail (thumbnailUrl, short-lived WebP, with thumbnailPlaceholderColor #rrggbb or null), and for a published destination the outcome of the steps after the publish, each a sentence or null: coverIssue, firstCommentIssue, captionIssue (YouTube subtitle languages that were not added, and why) and playlistIssue, plus firstCommentUrl (a link to a posted YouTube first comment). Requires the posts.read scope.
get_posting_limits
ChatGPTAnti-ban posting limits (cadence): pass socialAccountId for one account’s view (effective daily cap / spacing, used-today, reset instant), or omit it for the whole-tenant overview of every active channel plus the tenant bypass flag. Compare usedInWindow (not usedToday) with the cap: capWindow says how the platform counts it — utc-day, or rolling-24h where the platform counts the last 24 hours (TikTok, Instagram, Threads) — and nextSlotAt is when the next post may go out once the cap is full (null while there is room). appliedMinIntervalSeconds is the gap SkedCast holds between two posts right now — wider than effectiveMinIntervalSeconds while SkedCast is protecting the account; slowed says why (strike = the platform refused a post as spam or for posting too much, failure_spike = most recent posts failed, engagement_collapse = views fell toward zero) and until when (null = until the numbers recover); nextPostAt is the earliest the next post may go out under the cap, the gap and that protection (null = now).
get_quota
ChatGPTRead your adaptive-quota-pool budget for every scarce channel (today: YouTube uploads) BEFORE planning a batch: guaranteed slice, used, best-effort burst headroom, tomorrow’s guarantee tonight, and the window reset (ISO-8601 UTC; null for rolling windows). Read-only.
get_report
ChatGPTRead one report definition with its run history — each run’s status, window, and whether an export is available to download. Requires the reports.read scope.
get_search_keywords
ChatGPTThe search terms people used to find a Google Business Profile location, for one month (default the latest stored), refreshed nightly; also the months available. Use list_accounts to find the location’s socialAccountId. Requires the analytics.read scope and the deepAnalytics plan feature.
get_top_posts
ChatGPTThe window’s top posts by one figure (by: views, engagements, likes, comments, shares, saves, reach or engagementRate; default views), ranked over every post first published in the window — the same ranking the console’s Insights, Home and client report show. Each row is a per-post analytics row (as get_post_analytics returns). Narrow with clientId, platform or socialAccountId. Requires the analytics.read scope; narrowing to one account needs the deepAnalytics plan feature.
get_webhook
ChatGPTRead one webhook endpoint: its URL, subscribed events, active state, consecutive-failure count and, when it has been auto-disabled, when and why. The signing secret is NEVER returned here — it is shown once at creation and once on rotation. Requires the webhooks.read scope.
import_media
ChatGPTIngest an external image/video into the content library by URL — the file must already be reachable at a public URL. For bytes you already hold (an image your client just rendered, or a file on the caller’s own machine) use upload_media instead; this tool cannot accept file contents. This creates a real library asset — confirm with the user before calling. Requires the media.manage scope.
list_account_analytics
ChatGPTEvery connected channel’s analytics in one paged call: each item is what get_account_analytics returns for that channel (totals, daily series, profile activity, demographics, followsUniqueByDay on Facebook) plus its name, handle, client and whether it is disconnected. Newest-connected first; limit 1..25 (default 10); pass the returned nextCursor to page. Requires the analytics.read scope and the deepAnalytics plan feature.
list_accounts
ChatGPTList connected social accounts (with token health), optionally filtered by client / platform. An account that needs reconnecting carries reconnectReason, why: token_refresh_rejected, no_refresh_token, app_removed (SkedCast was removed on the platform), missing_permission, security_change (password or security change), credential_unreadable, repeated_auth_failures, moved_to_new_app, reconnect_requested or unknown (the cause could not be proven); absent otherwise. An account the customer removed from the console (it is disconnected, its history is kept) carries removedAt, an ISO timestamp; absent on every other account. An account with missingPermissions publishes, but its login predates permissions the platform has since approved (a Facebook Page connected before 2026-10-23 has no comments, first comments or Messenger): reconnecting adds them. Keyset-paginated: pass the returned nextCursor to page.
list_approvals
ChatGPTList the workspace's approval queue (pending and decided requests), newest-first, with optional status / post filters. Each row carries a preview of the gated post.
list_campaigns
ChatGPTList the workspace's campaigns (named post groupings), alphabetically, optionally filtered by client / status. Requires the campaigns.read scope.
list_clients
ChatGPTList the workspace’s client groups (brands / influencers), newest first. Keyset-paginated: pass the returned nextCursor to page.
list_library_assets
ChatGPTList media assets in the content library (DAM), with optional kind / folder / label / client / status / search filters. Newest-uploaded first. Keyset-paginated: pass the returned nextCursor to page. Each asset carries a short-lived WebP thumbnailUrl and placeholderColor (#rrggbb or null).
list_pinterest_board_sections
ChatGPTList the sections of one board on a connected Pinterest account, each as { id, name }. A section is optional: pass its id as boardSectionId beside the boardId to file the Pin under it. Requires the accounts.read scope.
list_pinterest_boards
ChatGPTList the boards a connected Pinterest account can pin to: its own boards and the group boards it collaborates on, each as { id, name }. Every Pin needs one: pass its id (digits only) as variants[].overrides.boardId on create_post, or per account as overrides.accountOverrides[accountId].boardId when a post goes to several Pinterest accounts (a board belongs to one account). A board name is accepted in place of the id only for a post to exactly one Pinterest account; with 2+ Pinterest accounts the name is refused (pinterest_board_unresolved), so pass the id from this list. Any other platform’s account is account_not_found. Requires the accounts.read scope.
list_post_targets
ChatGPTList individual post DESTINATIONS — one row per platform account a post goes to — with optional status / statuses / client / account / platform / free-text filters and a SCHEDULE-time window. This is the tool to ask "what publishes on this date": unlike list_scheduled_posts, from/to and sort=scheduled_at are about when a target GOES OUT, not when its post was authored. The page carries total (matching targets) and totalPosts (distinct parent posts) alongside nextCursor, so a caller can tell a full page from a complete set instead of assuming the default 25 rows are everything. A published row carries the outcome of the steps after the publish (coverIssue, firstCommentIssue, captionIssue, playlistIssue: a sentence or null each) and firstCommentUrl. A row on hold (status on_hold) carries holdReason — why it waits (platform_post_limit, spacing, slowed_platform_flag, channel_paused, x_budget, media_preparing, plan_limit, channel_needs_reconnect, …); a missed row (status missed: its time passed while it waited, reschedule it) carries outcomeReason (catch_up_window, plan_limit, channel_needs_reconnect); both are null otherwise. Every row also carries requiredAction — what you must do for it (reconnect_channel, change_plan, contact_support, reschedule or retry), or null when nothing is needed (a hold that lifts by itself, or any other status); act when it is not null. Keyset-paginated: pass the returned nextCursor to page.
list_reports
ChatGPTList the workspace’s saved analytics report definitions (CSV/PDF), newest first. Keyset-paginated: pass the returned nextCursor to page.
list_scheduled_posts
ChatGPTList the workspace’s composed posts (queue / drafts / published), most recently CREATED first, with optional status / client / platform / free-text filters and a creation-date window. Keyset-paginated: pass the returned nextCursor to page. Note the ordering and the from/to window are both on when a post was AUTHORED, not when it publishes — to reason about publish times, read each post’s targets. Each target shows its status, holdReason, outcomeReason and requiredAction (what you must do for it, or null).
list_webhook_deliveries
ChatGPTRead the delivery log for one endpoint — each attempt with its event type, response status, error and timing. This is how you find out WHY an endpoint stopped receiving: a 401 from your server, a timeout, or an auto-disable after repeated failures. Keyset-paginated: pass the returned nextCursor to page. Requires the webhooks.read scope.
list_webhooks
ChatGPTList the workspace's outbound webhook endpoints (url, subscribed events, active flag; signing secret masked), newest-first. Requires the webhooks.read scope.
publish_draft
ChatGPTPublish or schedule an existing DRAFT post in place: pass the draft id plus the full post body (the same arguments as update_post). schedule decides now or later. Runs the same checks as create_post — platform validation, the client's approval policy, quota and pacing — so a client that requires approval gets the post submitted for review instead. The post keeps its id. Refuses a post that is no longer a draft or is waiting for review. Retry-safe: identical arguments replay the original result. This puts real content on connected accounts — confirm with the user before calling. Requires the posts.compose scope.
reschedule_post
ChatGPTMove a pending post target to a new time — provide an absolute scheduledAt (ISO-8601 with offset) or a relative shiftMinutes. This changes a real scheduled post — confirm with the user before calling. Requires the posts.compose scope.
retry_post
ChatGPTRe-queue a post target so it publishes again immediately: a failed one, or one on hold whose hold reason may be cut short (platform_rate_limit, platform_outage, shared_capacity, quota_unconfirmed). Any other status or hold reason is refused (409 not_retryable); a missed target is rescheduled, not retried. This re-publishes a real post — confirm with the user before calling. Requires the posts.compose scope.
revise_post
ChatGPTChange the CONTENT of an already-scheduled post — caption, media, cover, alt text, per-platform overrides — without disturbing its destinations. Nothing about the fan-out is rewritten: the publisher reads content at publish time, so every destination picks the change up on its own schedule. Send only the fields you are changing; an omitted field is left alone, and variants are merged by platform. Destinations and schedule are NOT editable here — use reschedule_post / cancel_post / bulk_post_target_action for those. REFUSES, naming the destinations that blocked it, when any has already published or has been handed to the publisher. If the post was already APPROVED, the edit returns the whole post to review and nothing publishes until it is approved again — the result's reGated says so. A YouTube video that is already live is changed with edit_published_post, not here. The result carries warnings for the revised overrides. A write action — confirm with the user before calling. Requires the posts.compose scope.
rotate_webhook_secret
ChatGPTMint a NEW signing secret for one endpoint and return it. THE SECRET IS SHOWN ONCE and is not readable again — hand it to the user, and do not repeat it in summaries, logs or files. The previous secret keeps verifying for a short overlap so the customer can redeploy without dropping deliveries. A write action — confirm with the user before calling. Requires the webhooks.manage scope and the webhooks feature.
set_post_recycle
ChatGPTSet, update, or disable a post's evergreen recycle settings (enabled / intervalDays / times / endDate / maxRecycles) — the shared recycle-config contract. Pass enabled:false to pause recycling without losing the settings. This changes how a real post republishes — confirm with the user before calling. Requires the posts.compose scope.
shift_campaign
ChatGPTMove a whole campaign's still-pending schedule by a relative shiftMinutes (non-zero, ±1 year) — every post the reschedule engine can move, including posts held by the plan; posts waiting for a reviewer are reported as skipped, not moved. Returns the accepted / skipped counts. This moves real scheduled posts in bulk — confirm with the user before calling. Requires the campaigns.manage scope.
test_webhook
ChatGPTSend a webhook.test delivery to one endpoint and report the outcome, so you can prove an endpoint is reachable and correctly verifying signatures WITHOUT waiting for a real event. This performs a real outbound HTTP request to the customer’s URL — confirm with the user before calling. Requires the webhooks.manage scope and the webhooks feature.
unassign_from_campaign
ChatGPTRemove posts from a campaign. Grouping only — the posts themselves and their schedules are untouched. A write action — confirm with the user before calling. Requires the campaigns.manage scope.
update_campaign
ChatGPTChange a campaign’s name, description, status, colour or date range. At least one field is required. A write action — confirm with the user before calling. Requires the campaigns.manage scope.
update_client
ChatGPTChange a client’s name, type, approval policy, posting rules, brand voice or default timezone. A write action — confirm with the user before calling. Requires the clients.manage scope.
update_post
ChatGPTReplace a DRAFT post in place — its content, media, targets, and schedule. Only draft posts can be updated (a scheduled or published post returns a conflict). This rewrites a real draft — confirm with the user before calling. Requires the posts.compose scope.
update_webhook
ChatGPTChange an endpoint’s URL, its subscribed event list, or whether it is active. Re-enabling an endpoint that was auto-disabled after repeated failures is done here. The event list REPLACES the previous one — send the full set, not a delta. A write action — confirm with the user before calling. Requires the webhooks.manage scope and the webhooks feature.
upload_media
ChatGPTUpload an image or video from your own machine into the content library, and get back a media id usable as create_post's mediaIds. Three phases: inline sends the file base64 in this call (≤18 MiB, works in any client); start returns a presigned URL your client PUTs the bytes to directly (any size, needs a client that can make HTTP requests — e.g. a shell); finish registers what start uploaded. For a file already on the public internet use import_media instead. This creates a real library asset — confirm with the user before calling. Requires the media.manage scope.
validate_post
ChatGPTDry-run a compose request and get the per-platform verdict WITHOUT writing anything: same arguments as create_post, same validation gate, no post row, no draft, no idempotency record. Returns { valid, platforms[], estimatedCostUsd, verdicts? }, where each platform carries its own valid plus the issues that would block it. It also returns warnings for a YouTube override key that is unknown, misplaced or deprecated (see create_post). Use this instead of creating a far-future post and cancelling it. Note it does NOT run cadence or quota admission, so valid: true means the SHAPE is acceptable, not that a publish cannot later be deferred. A Pinterest board value create_post would refuse (a name on a post to 2+ Pinterest accounts, a name no board has) is refused here with the same pinterest_board_unresolved error, and a name for one Pinterest account is checked as the board id it resolves to. Requires the posts.compose scope.
assign_to_campaign
Claudebulk_post_target_action
Claudecancel_post
Claudeclear_post_recycle
Claudecreate_campaign
Claudecreate_client
Claudecreate_connect_invite
Claudecreate_post
Claudecreate_posts_bulk
Claudecreate_webhook
Claudedecide_approval
Claudedelete_campaign
Claudedelete_client
Claudedelete_webhook
Claudedownload_report_export
Claudeget_account
Claudeget_account_analytics
Claudeget_agency
Claudeget_analytics_overview
Claudeget_approval
Claudeget_best_times
Claudeget_campaign
Claudeget_capabilities
Claudeget_client
Claudeget_client_analytics
Claudeget_media_asset
Claudeget_post
Claudeget_post_analytics
Claudeget_post_recycle
Claudeget_post_target
Claudeget_posting_limits
Claudeget_quota
Claudeget_report
Claudeget_webhook
Claudeimport_media
Claudelist_accounts
Claudelist_approvals
Claudelist_campaigns
Claudelist_clients
Claudelist_library_assets
Claudelist_post_targets
Claudelist_reports
Claudelist_scheduled_posts
Claudelist_webhook_deliveries
Claudelist_webhooks
Claudepublish_draft
Claudereschedule_post
Clauderetry_post
Clauderevise_post
Clauderotate_webhook_secret
Claudeset_post_recycle
Claudeshift_campaign
Claudetest_webhook
Claudeunassign_from_campaign
Claudeupdate_campaign
Claudeupdate_client
Claudeupdate_post
Claudeupdate_webhook
Claudeupload_media
Claudevalidate_post
Claude