get_account_metrics
ChatGPTAccount-level performance metrics for a workspace's own social accounts — engagements, reach, views, content counts, EMV, follower metrics, and more. Returns one row per (metric, platform), with total_value, average_value, and a values series of {end_time, value} buckets. tick_size defaults to DAY when omitted; timezone defaults to UTC. This tool is account-level only — it cannot explain why a specific post or format performed the way it did; use search_owned_content for per-post detail. Not paginated; scope metrics and platforms to what you need, since requesting the full set over a long window returns a large payload. Queries the workspace this connection is currently in unless you pass a workspaceId; call get_available_workspaces first to get a valid id for a different one.
get_available_accounts
ChatGPTList the social accounts connected to a workspace — id, workspace, handle, platform and profile picture — filterable by platform and name, with paging. Queries the workspace this connection is currently in unless you pass a workspaceId; call get_available_workspaces first to get a valid id for a different one.
get_available_competitor_groups
ChatGPTList a workspace's saved competitor groups — id, name, and the competitor ids each one contains. Members are narrowed to the competitors the workspace still actively tracks, so every id also appears in get_available_competitors. Ordered oldest first. Takes no filters and no paging; an empty list means the workspace has saved no groups. The built-in "all competitors" scope is not a group — reach it by omitting the competitor filters entirely. Queries the workspace this connection is currently in unless you pass a workspaceId; call get_available_workspaces first to get a valid id for a different one.
get_available_competitors
ChatGPTList the competitors a workspace actively tracks — id, brand name, profile picture, the Instagram, TikTok, X and YouTube handles, and created_at. Ordered by brand name. Takes no filters and no paging: you always get the whole tracked set, and an empty list when the workspace tracks nobody. Queries the workspace this connection is currently in unless you pass a workspaceId; call get_available_workspaces first to get a valid id for a different one.
get_available_labels
ChatGPTList the labels defined in a workspace — id, title, and creation date — filterable by title, with paging. Queries the workspace this connection is currently in unless you pass a workspaceId; call get_available_workspaces first to get a valid id for a different one.
get_available_products
ChatGPTList the products in a workspace — id, title, source (SHOPIFY or CUSTOM), and creation date — filterable by source and title, with paging. Queries the workspace this connection is currently in unless you pass a workspaceId; call get_available_workspaces first to get a valid id for a different one.
get_available_workspaces
ChatGPTList the workspaces this user can access — name, id, and isActive flag. The entry with isActive=true is the workspace the current connection is in; every tool defaults to that workspace when no workspaceId is given. Use search for a case-insensitive partial match, or set fuzzySearch=true when the user's wording may be approximate or have typos.
get_competitor_comparison
ChatGPTCompare competitors side by side over a date window on the metrics you choose. metrics is required — pick from VOLUME, ENGAGEMENTS, ENGAGEMENTS_PER_POST, LIKES, COMMENTS, SHARES, VIEWS_ESTIMATE, EMV, AVG_ENGAGEMENT_RATE, SHARE_OF_VOICE, SENTIMENT, AUDIENCE_GROWTH. Most come back as a plain number; AVG_ENGAGEMENT_RATE and SHARE_OF_VOICE as percentages; VOLUME, ENGAGEMENTS, VIEWS_ESTIMATE, EMV, and AUDIENCE_GROWTH come back as {total, by_platform} where platforms with a zero value are omitted; SENTIMENT comes back as a net promoter score plus sentiment-bucket counts and percentages. AUDIENCE_GROWTH is the competitor's follower count at the latest snapshot in the window, not a change over it. SHARE_OF_VOICE and SENTIMENT ignore the ownership filter that every other metric honors. Returns one row per competitor — id, brand name, and the requested metrics. Every competitor in scope comes back, zero-filled if it published nothing in the window; returns an empty list when none are in scope. competitor_ids and competitor_group_id are mutually exclusive; omit both to cover every competitor the workspace tracks. The window defaults to the last 30 days, may span at most 365 days, and covers posts published from start_date up to but not including end_date. Queries the workspace this connection is currently in unless you pass a workspaceId; call get_available_workspaces first to get a valid id for a different one.
get_competitor_summary
ChatGPTGet headline metrics for the tracked competitors over a date window: total mentions, estimated views, engagements, engagements per post, unique authors and positive-sentiment share — once overall, then per competitor with its brand name and earned share of voice. Mentions, unique authors, sentiment, and share of voice count earned posts only (posts the competitor did not publish itself) and ignore the ownership filter; views, engagements, and engagements per post honor it. Every competitor in scope comes back, zero-filled if it had nothing in the window; when none are in scope the overall numbers are zeros and the per-competitor list is empty. competitor_ids and competitor_group_id are mutually exclusive; omit both to cover every competitor the workspace tracks. The window defaults to the last 30 days, may span at most 365 days, and covers posts published from start_date up to but not including end_date. Queries the workspace this connection is currently in unless you pass a workspaceId; call get_available_workspaces first to get a valid id for a different one.
get_competitor_trend
ChatGPTTrack competitor metrics over time, bucketed by day, week, or month (tick_size, defaults to DAY). metrics is required — pick from ENGAGEMENTS, VIEWS_ESTIMATE, VOLUME, EMV, AUDIENCE_GROWTH. Returns one series per (metric, competitor): the four content metrics are totaled across platforms into a single line per competitor, and AUDIENCE_GROWTH is the competitor's combined follower count across platforms at each snapshot point; AUDIENCE_GROWTH ignores the ownership filter that the other metrics honor. Each series is labeled CROSS_PLATFORM unless the request filters to exactly one platform, in which case that platform's name labels it instead. Empty series when no tracked competitors are in scope. competitor_ids and competitor_group_id are mutually exclusive; omit both to cover every competitor the workspace tracks. The window defaults to the last 30 days, may span at most 365 days, and covers posts published from start_date up to but not including end_date. Queries the workspace this connection is currently in unless you pass a workspaceId; call get_available_workspaces first to get a valid id for a different one.
get_owned_content_summary
ChatGPTGet aggregate performance of a workspace's own published posts, instead of the posts themselves — an overall bucket plus one per platform (Instagram, Facebook, TikTok, X, YouTube, LinkedIn, Pinterest). Each bucket carries the content count, totals (comments, reactions, shares, saves, views, engagement, EMV, time watched), per-post averages of the same plus average reach and video duration, and rates as percentages (engagement rate by views with and without stories, full-video / three-second / six-second watch rates). Values are null on platforms that don't report that metric. category defaults to ORGANIC — pass BOOSTED or AD to aggregate those instead. Posts whose metrics have not been fetched yet are left out; an empty match returns zeroed buckets. Accepts the same filters as search_owned_content minus sorting and paging. The window defaults to the last 365 days; the resolved start and end dates come back in the response. Queries the workspace this connection is currently in unless you pass a workspaceId; call get_available_workspaces first to get a valid id for a different one.
search-inbox-items
ChatGPTSearch the workspace's unified engagement feed: comments and DMs on its own content and tagged UGC, plus reviews and untagged brand mentions. An untagged UGC post appears here as an UNTAGGED_BRAND_MENTION item, but the comments under it do not. It runs the same search as the web inbox, so it returns the same items in the same order. searchQuery matches comment, message, note, and author-handle text. Every token must appear as a whole word, so "cat" does not match "category". searchMode narrows that match to COMMENTS, DM, NOTES, or USER. As in the web inbox, a searchQuery ignores tab, accountIds, spamView, and every filter. To filter, omit searchQuery. Results come in pages of limit items. When hasMore is true, call again with page + 1 and the same inputs.
search_competitor_content
ChatGPTSearch the posts behind the competitive-listening numbers — caption, permalink, platform, format, surface, publish date, sentiment, AI summary, ownership, metrics (likes, comments, shares, estimated views, engagement rate, EMV), the creator who posted it (handle, profile picture, follower count, verified) and its media. One entry per post and no competitor id, so a post mentioning several tracked competitors appears once. Posts with no engagements and no views are left out, and Reddit posts are anonymized — blank caption, author handle and picture, zero follower count, no media. Sortable by DATE, ENGAGEMENT, ENGAGEMENT_RATE, LIKES, VIEWS, COMMENTS, SHARES (newest first by default); sort=ENGAGEMENT with a small limit returns the most viral posts. When no competitors are in scope, total is 0 and the page is empty. The window defaults to the last 30 days, may span at most 365 days, and covers posts from published_after up to but not including published_before. Queries the workspace this connection is currently in unless you pass a workspaceId; call get_available_workspaces first to get a valid id for a different one. Call get_available_competitors for competitor ids, or get_available_competitor_groups for a competitor_group_id — pass one or the other, never both. Omit both to cover every tracked competitor. IMPORTANT: This tool's results also render automatically as a visual card row (when the channel supports it) — every eligible item on the page becomes a card, regardless of how many you actually include in your text response or table. Set limit to exactly the number of items you intend to present, never a larger "buffer" amount.
search_owned_content
ChatGPTSearch the posts a workspace published from its own connected accounts — caption, permalink, platform, surface, media type, publish date, whether the post is organic, boosted or an ad, the posting account, media, applied labels, and per-post metrics (comments, reactions, shares, saves, reach, views, engagement, engagement rate by views as a percentage, EMV). Every metric is the combined organic-plus-paid total, and is null on platforms that don't report it. Filterable by account, surface, media type, category, label and publish date, with sorting and paging. category defaults to ORGANIC — pass BOOSTED or AD to see those instead. Posts whose metrics have not been fetched yet are left out, so very recent posts can be missing. Check meta.has_next to page through results. Queries the workspace this connection is currently in unless you pass a workspaceId; call get_available_workspaces first to get a valid id for a different one.
get_account_metrics
Claudeget_available_accounts
Claudeget_available_competitor_groups
Claudeget_available_competitors
Claudeget_available_labels
Claudeget_available_products
Claudeget_available_workspaces
Claudeget_competitor_comparison
Claudeget_competitor_summary
Claudeget_competitor_trend
Claudeget_owned_content_summary
Claudesearch_competitor_content
Claudesearch_owned_content
Claude