add_competitors_to_category
ChatGPTAdd monitored competitors to an existing competitor category. Current members are kept; nothing is removed. A competitor can be in several categories. WRITE ACTION. - competitor_category_id: from list_competitor_categories. - monitored_competitor_ids: 1 to 200 ids from list_monitored_competitors. Unknown ids reject the whole call. Returns added_competitor_ids, already_in_category_ids and the category's full monitored_competitor_ids after the change.
add_prompts_to_group
ChatGPTAdd monitored prompts to an existing prompt group. Prompts already in the group are kept as they are; nothing is removed. WRITE ACTION. A prompt can belong to several groups. - group_id: from list_groups. - prompt_ids: 1 to 200 ids from list_monitored_prompts. Unknown ids reject the whole call. Returns added_prompt_ids, already_in_group_ids and the group's prompts_count.
create_competitor_category
ChatGPTCreate a competitor category (a user-defined label such as 'Direct' or 'Regional' used to segment monitored competitors), optionally already containing competitors. WRITE ACTION — confirm the category name with the user before calling. - name: up to 60 characters, unique within the brand (case-insensitive); a duplicate fails with competitor_category_already_exists. - description (optional): up to 500 characters. - color (optional): badge color in the app, default "blue". - monitored_competitor_ids (optional): ids from list_monitored_competitors. Unknown ids reject the whole call. Returns {id, name, description, color, created_at, monitored_competitor_ids}. The id works as category_ids in get_dashboard, get_monitored_prompt and list_monitored_competitors.
create_monitored_competitor
ChatGPTStart tracking a new competitor alongside a brand, optionally from a competitor suggestion and already placed in competitor categories. WRITE ACTION — confirm the competitor name (and website and categories) with the user before calling. Counts against the workspace plan's monitored-competitor limit (error monitored_competitors_limit_reached when full; see get_usage_limits). - competitor_suggestion_id (optional): id from list_competitor_suggestions. Recommended when the competitor is in that list: the mentions already collected under that name count for the new competitor. Without name, the suggestion's name is used. - name: required when there is no competitor_suggestion_id. The competitor brand name exactly as AI models write it — mentions are matched by name. Unique within the brand (case-insensitive). - website (optional): full URL, e.g. https://example.com. - category_ids (optional): ids from list_competitor_categories; the new competitor is added to all of them. Any invalid suggestion or category id rejects the whole call and nothing is created. Mentions in past responses are recalculated in the background (calculation_status starts as "pending"); the competitor shows up in dashboard rankings once that finishes. Returns {id, name, website, calculation_status, competitor_suggestion_id, category_ids}.
create_monitored_prompts
ChatGPTStart monitoring new prompts (questions tracked across AI models) for a brand. WRITE ACTION — confirm the exact prompt texts with the user before calling. Every new prompt counts against the workspace plan's monitored-prompt limit and is immediately queued for its first run on every AI model. - prompts: one or more prompt texts, written the way a real user would ask an AI assistant (in the brand's language). There is no per-call cap; the only bound is the plan limit — call get_usage_limits first to know how many prompts still fit. - group_ids (optional): ids from list_groups; every new prompt is added to all of them. Unknown ids reject the whole call. The batch is all-or-nothing against the plan limit: if it does not fit, no prompt is created and the error carries used/limit. Prompts identical (case-insensitive) to one already monitored in the brand are skipped and returned in skipped_duplicates. Returns created [{id, prompt, group_ids}], skipped_duplicates and the workspace's monitored_prompts_usage {used, limit} after the call. Results for new prompts appear in the read tools after their first runs complete.
create_prompt_group
ChatGPTCreate a prompt group (a thematic category of monitored prompts, e.g. 'Product Comparisons') in a brand, optionally already containing prompts. WRITE ACTION — confirm the group name with the user before calling. - name: unique within the brand (case-insensitive). If a group with that name exists, the call fails with group_already_exists and its id. - prompt_ids (optional): monitored prompt ids from list_monitored_prompts. Unknown ids reject the whole call. Returns {id, name, prompt_ids}. The id works as group_id in the read tools and in add_prompts_to_group / remove_prompts_from_group.
get_brand
ChatGPTGet full details about a specific brand including its description, key features, website, country, and language. Use this to understand what the brand is and what it offers before diving into visibility analysis. The description and key_features fields provide context that helps interpret mention rates and sentiment in other tools' output.
get_cited_source_related_prompts
ChatGPTList which monitored prompts have AI responses that cite a specific domain. This reveals the connection between a web source and the questions it influences. Requires domain_id obtained from list_cited_sources. Returns a paginated list of prompts that reference this domain, allowing you to understand: - Which types of questions lead AI models to cite this source - How broadly influential a domain is across different prompt categories - Whether a competitor's website is cited in prompts relevant to your brand Combine with get_monitored_prompt_responses to read the actual AI responses that cite this domain and understand the context of each citation.
get_cited_source_url_related_prompts
ChatGPTList which monitored prompts have AI responses that cite a specific URL. This is the most granular level of source-to-prompt mapping. Requires domain_id from list_cited_sources and url_id from get_cited_source_urls. Returns a paginated list of prompts where AI responses reference this exact URL. Use this to understand: - Exactly which questions lead AI models to cite a particular page - The scope of influence of a single piece of content on AI answers - Whether a specific competitor article is affecting multiple prompts This is the deepest drill-down in the cited sources hierarchy: list_cited_sources → get_cited_source_urls → get_cited_source_url_related_prompts
get_cited_source_urls
ChatGPTList specific URLs (pages) from a cited domain, with detailed impact metrics. This drills down from domain-level to page-level citation analysis. Requires domain_id obtained from list_cited_sources. Returns a paginated list where each URL includes: - id: URL identifier (use as url_id in get_cited_source_url_related_prompts) - url: the full page URL - impact: impact score for this specific page - citation_count: how many times this URL was cited - prompts_count: how many different prompts reference this URL - first_seen_at: when this URL first appeared in AI citations - ai_models: which AI platforms have cited this URL Use 'search' to filter URLs by keyword (e.g. search='review' to find review pages). This tool is valuable for: - Identifying the most influential individual pages within a domain - Finding specific articles or pages that AI models rely on - Prioritizing content outreach targets (which pages to get mentioned on) - Tracking when new URLs start appearing in AI citations
get_competitor_category
ChatGPTGet a competitor category and the monitored competitors in it. Returns: - data: id, name, description, color, created_at, updated_at, and competitors — one page of the category's competitors, each with id, name, website, created_at and categories (all categories it belongs to, as [{id, name}]) - pagination: current_page, per_page, total_pages, total_records — refers to data.competitors (100 per page by default, max 100). Increase page to fetch the remaining competitors. Get competitor_category_id from list_competitor_categories.
get_dashboard
ChatGPTGet the brand's high-level dashboard with aggregated metrics across all monitored prompts. This is the best starting point for a quick brand health overview and competitive positioning assessment. Returns: - total_prompts: number of prompts being monitored - total_responses: total AI responses collected in the date range - total_brand_mentions: how many responses mentioned the brand - brand_mention_rate: percentage of responses mentioning the brand (0-100) - average_brand_sentiment: overall sentiment toward the brand (-1 to 1) - competitors_count: number of tracked competitors - brands_ranking: competitive leaderboard showing the brand and all competitors ranked by mention rate — this is the key competitive intelligence metric - top_ranked_prompts: prompts where the brand performs best - lowest_ranked_prompts: prompts where the brand performs worst (these are priority targets for optimization) Use group_id to filter metrics to a specific prompt category. Use ai_model to see performance on a single AI platform. This tool provides aggregated summary data. For deeper analysis: - Use list_monitored_prompts to see per-prompt breakdown - Use get_monitored_prompt on lowest_ranked_prompts to investigate issues - Use list_monitored_competitors to identify who the competitors are TIP: Call this tool with different ai_model values to compare brand performance across AI platforms (e.g. ChatGPT vs Perplexity vs Gemini).
get_monitored_competitor
ChatGPTGet full details about a specific monitored competitor, including its name, description, and identification metadata. Use this when you need more context about a particular competitor that appears in dashboard rankings or prompt-level analysis.
get_monitored_prompt
ChatGPTGet detailed statistics and daily time-series data for a specific monitored prompt. This is the go-to tool for deep-diving into a single prompt's performance over time, including competitor comparison. Returns: - Overall metrics: total_responses, brand_mentions, mention_rate, brand_average_position, average_sentiment - Daily time-series: for each day in the date range, shows mention rates and sentiment for the brand AND each of its competitors, enabling trend analysis and competitive benchmarking over time This tool is essential for: - Tracking how brand visibility changes day-by-day for a specific query - Comparing the brand vs competitors on a specific prompt - Identifying trends (improving or declining mention rates) - Spotting anomalies (sudden drops or spikes in visibility) To get the prompt text and summary stats first, use list_monitored_prompts. To read the actual AI-generated responses, follow up with get_monitored_prompt_responses.
get_monitored_prompt_citations
ChatGPTGet domains and URLs that AI models cited in their responses to a specific monitored prompt. This reveals which web sources influence how AI models answer this particular question. Returns a paginated list of domains ranked by citation frequency, where each domain includes: - domain name and citation count - list of specific URLs cited from that domain Use this to understand: - Which websites are most influential for this specific query - Whether the brand's own website is being cited - Which competitor or third-party sites are shaping AI responses - Potential content optimization targets (sites to get listed on) For a broader cross-prompt view of cited sources, use list_cited_sources instead. For the brand-level source analysis, combine this tool's output across multiple prompts of interest.
get_monitored_prompt_responses
ChatGPTGet the actual full-text AI-generated responses for a monitored prompt. This is the only tool that provides the raw content of what AI models actually said, making it critical for qualitative analysis. Returns a paginated list where each response includes: - ai_model: which AI platform generated this response (e.g. chat_gpt, perplexity) - response_markdown: formatted response with markdown - response_text: plain text version of the response - brand_mentioned: whether the brand was mentioned (true/false) - brand_sentiment: sentiment toward the brand in this response (-1 to 1) - citations: list of URLs the AI model cited in its answer - search_queries: search queries the AI model executed to build its answer - created_at: when this response was captured Use this tool to: - Read exactly what AI models say about the brand - Understand the context in which the brand is (or isn't) mentioned - Analyze how different AI models frame the brand differently - Identify specific claims or inaccuracies in AI responses - See which sources are cited alongside brand mentions NOTE: Responses can be lengthy. For prompts with many responses, you may need multiple paginated calls. Filter by ai_model to focus on one platform.
get_monitored_prompt_search_queries
ChatGPTGet the search queries that AI models internally executed when building their response to a monitored prompt. Many AI models perform web searches before generating answers — this tool reveals what they searched for. Returns a paginated list where each entry includes: - query: the search query string - count: how many times this query was used across responses Results are sorted by frequency (most common queries first). Use this to understand: - What information AI models seek when answering this prompt - Whether AI models search for the brand by name - What related terms or competitor names AI models look up - SEO and content strategy insights (what queries to optimize for) This data is especially valuable for content optimization: if AI models consistently search for specific terms, ensuring your brand ranks well for those terms can improve AI visibility.
get_recommendation
ChatGPTGet one recommendation in full, including its payload — the actual instructions/content needed to execute it. Returns everything from list_recommendations plus: - payload: the body of the recommendation. IMPORTANT: the keys of this object DEPEND ON 'kind' — each kind has its own schema (an llms_txt recommendation carries a drafted file, a new_article one carries the article brief, a reddit_comment one carries the post and a suggested comment, and so on). Read the keys that are actually present; do not assume a fixed shape. - payload_schema_version: schema version of that payload - dismiss_reason / dismiss_reason_code / dismissed_at: why it was dismissed, when applicable See list_recommendations for what each 'status' value means — in particular, 'done' may have been resolved automatically by FirstAnswer rather than by the user, and 'dismissed' is a lasting user opt-out, not pending work. Use list_recommendations first to discover recommendation_id values.
get_usage_limits
ChatGPTShow plan usage and limits: how many monitored prompts and monitored competitors each workspace uses versus what its plan allows, and how many MCP tool calls are left this billing cycle. This tool is FREE: it never counts against the MCP tool call quota. Call it before create_monitored_prompts / create_monitored_competitor to know how many still fit (the limits are per workspace, shared by all its brands, and come from the workspace owner's plan), or whenever the user asks about their plan usage. - brand_id (optional): only the workspace that brand belongs to. Without it, every workspace the user can reach through MCP is returned. Returns workspaces [{workspace_id, workspace_name, plan_name, brand_ids, monitored_prompts {used, limit, remaining}, monitored_competitors {used, limit, remaining}}] and mcp_tool_calls {metered, used, limit, remaining}. metered=false means the plan includes unlimited MCP access (used/limit are null); metered=true means a monthly complimentary quota applies.
link_competitor_suggestions
ChatGPTLink suggested names to a competitor that is ALREADY monitored, so the mentions of those names count for it — e.g. "Rival Inc" or "Rival CRM" for the monitored competitor "Rival" (variants, product names, old names). The monitored competitor keeps its own name. WRITE ACTION — confirm with the user which names belong to which competitor before calling. - monitored_competitor_id: from list_monitored_competitors. - competitor_suggestion_ids: 1 to 200 ids from list_competitor_suggestions. Unknown ids, or names already counted for ANOTHER monitored competitor, reject the whole call (error details say which competitor owns them). Names already linked to this competitor are skipped. The linked names leave list_competitor_suggestions, and the competitor's mentions are recalculated in the background (calculation_status "pending" until it finishes). Returns linked_suggestion_ids, already_linked_ids, linked_names (every name now counted for the competitor) and calculation_status.
list_brands
ChatGPTList all brands available in the user's workspaces. This is the ENTRY POINT for all analyses — call this first to obtain brand_id values needed by every other tool. Returns a paginated list where each brand includes: - id: unique brand identifier (use as brand_id in other tools) - workspace_id: workspace this brand belongs to (null if legacy / unassigned) - workspace_name: human-readable workspace name - name: internal brand name - display_name: human-friendly brand name for display - description: what the brand/company does - key_features: notable products or services - website: brand's primary URL - country / language: market context - created_at: when monitoring started If the user has multiple brands, you may need to iterate pages or ask which brand they want to analyze.
list_cited_sources
ChatGPTList all domains (websites) cited by AI models in their responses across ALL monitored prompts for a brand. This provides a brand-wide view of which web sources are influencing AI-generated answers. Returns a paginated list where each domain includes: - id: domain identifier (use as domain_id in related tools) - domain: the domain name (e.g. 'wikipedia.org', 'forbes.com') - impact: impact score representing how influential this source is - prompts_count: how many different prompts cite this domain - responses_count: total responses that cite this domain - urls_count: number of distinct URLs cited from this domain - ai_models: list of AI platforms that cited this domain Use 'search' to filter by domain name (e.g. search='wikipedia'). Use group_id to narrow to citations from a specific prompt category. This is the starting point for source analysis. To drill deeper: 1. Call get_cited_source_urls with domain_id to see specific pages cited 2. Call get_cited_source_related_prompts to see which prompts reference this domain 3. Combine with get_monitored_prompt_citations for per-prompt source breakdown Identifying top cited sources helps understand what content AI models trust and where the brand should aim to be referenced.
list_cited_urls
ChatGPTList all individual URLs cited by AI models across all domains and prompts for a brand. Unlike get_cited_source_urls which drills into a single domain, this tool returns a flat, cross-domain ranking of every cited URL. Returns a paginated list where each URL includes: - url_id: URL identifier - url: the full page URL - domain_id: identifier of the parent domain - domain: the domain name (e.g. 'wikipedia.org') - ai_models: list of AI platforms that cited this URL - impact: percentage of total responses that cited this URL (0-100) - count: total number of times this URL was cited - prompts_count: how many different monitored prompts cite this URL - first_seen_at: when this URL first appeared in AI citations Use 'search' to filter by URL substring. Use monitored_prompt_id to narrow results to URLs cited in a specific prompt's responses. This tool is valuable for: - Getting a global view of the most influential pages across all domains - Finding which specific pages AI models rely on most - Discovering new URLs entering AI citations (sort by first_seen_at) - Comparing URL-level influence across AI platforms
list_competitor_categories
ChatGPTList the competitor categories of a brand. Categories are user-defined labels that group monitored competitors (e.g. "Direct", "Regional", "Priority"); a competitor can belong to several categories. Returns a paginated list where each category includes: - id: category identifier (use as competitor_category_id in get_competitor_category, and in category_ids to filter get_dashboard, get_monitored_prompt and list_monitored_competitors) - name: category display name - description: what the category means for this brand (may be null) - color: color slug used in the app - competitors_count: number of monitored competitors in the category - created_at: when the category was created
list_competitor_suggestions
ChatGPTList competitor suggestions: brands that AI models mention in this brand's monitored responses but that are NOT monitored competitors yet — the same list the app shows on the competitors screen. Ordered by times_mentioned (most mentioned first). Brands mentioned only sporadically are left out (the cut-off grows with the number of candidates), and suggestions the user chose to ignore in the app are not included. Each item: id (pass it as competitor_suggestion_id to create_monitored_competitor so the mentions already collected for that name count for the new competitor), name (as the AI models wrote it) and times_mentioned. Paginated like the other list tools (page, per_page max 100; pagination: current_page, per_page, total_pages, total_records). Typical flow: list_competitor_suggestions → pick the relevant ones with the user → list_competitor_categories → create_monitored_competitor with competitor_suggestion_id and category_ids.
list_groups
ChatGPTList prompt groups (categories) for a brand. Groups organize monitored prompts into thematic categories such as 'Product Comparisons', 'Industry Questions', or 'Purchase Intent'. Returns each group with: - id: group identifier (use as group_id filter in other tools) - name: group display name Use group_id to filter results in list_monitored_prompts, list_cited_sources, and get_dashboard to narrow analysis to a specific category of prompts. This is especially useful when the brand has many monitored prompts and you want to focus on a subset.
list_monitored_competitors
ChatGPTList all competitors being tracked alongside a brand. Competitors are other brands/companies whose mentions in AI responses are tracked for comparison with the main brand. Returns a paginated list where each competitor includes: - id: competitor identifier - name: competitor brand name - display_name: human-friendly name - description: what the competitor does Use this to understand the competitive landscape before analyzing per-prompt data. Competitor names will appear in the brands_ranking from get_dashboard and in per-prompt breakdowns from get_monitored_prompt. Call this tool to map competitor IDs to names when interpreting those results.
list_monitored_prompts
ChatGPTList all monitored prompts for a brand with aggregated performance statistics over the specified date range. This is the primary tool for understanding which questions/queries are being tracked and how the brand performs in each one. Returns a paginated list where each prompt includes: - id: prompt identifier (use as monitored_prompt_id in detail tools) - text: the actual prompt/question being monitored - group: which category the prompt belongs to - brand_mentions: total number of times the brand was mentioned - brand_average_position: average ranking position when mentioned (lower = better) - mention_rate: percentage of responses that mention the brand (0-100) - average_sentiment: sentiment score for brand mentions (-1 to 1) - response_count: total AI responses collected Use group_id to filter by prompt category. Use ai_model to see performance on a specific AI platform. IMPORTANT: This returns summary statistics only. For deeper analysis of any prompt, follow up with: - get_monitored_prompt: time-series data and competitor comparison - get_monitored_prompt_responses: actual AI response texts - get_monitored_prompt_citations: which sources AI models cited - get_monitored_prompt_search_queries: what search queries AI models performed Dates must be in YYYY-MM-DD format.
list_recommendations
ChatGPTList the brand's recommendations — the concrete, prioritized work FirstAnswer suggests to improve the brand's AI visibility (technical fixes, content gaps, setup steps, community engagement). This is the same list the user sees under "Recommendations" in the FirstAnswer app. Returns a paginated list where each recommendation includes: - id: recommendation identifier (pass it to get_recommendation for the full body) - category: on_site_technical | content_gap | setup | community - kind: the specific recommendation type. The full set is 'robots_txt_blocked_ai_crawler', 'llms_txt', 'new_article', 'analytics_script_setup', 'crawler_logs_setup', 'reddit_comment' - title: human-readable summary of what to do - priority: low | medium | high - status: see the status meanings below - completed_at, created_at, updated_at STATUS MEANINGS (important — these are not generic workflow states): - 'suggested': FirstAnswer detected the opportunity and is proposing it. This is the state every recommendation starts in, and the state a previously completed one returns to if the problem comes BACK (a regression, e.g. a robots.txt rule that was fixed and later reappeared). The user has not decided on it yet. - 'approved': the user reviewed it and accepted it — they intend to do it, but it is NOT done yet. Still outstanding work. - 'done': completed. Either the user marked it done, or FirstAnswer detected on a later check that the problem no longer exists and resolved it automatically. Do not assume a human did the work. - 'dismissed': the user explicitly rejected it (dismiss_reason_code says why: not_relevant, not_priority, disagree, other). This is a lasting opt-out — FirstAnswer will not propose it again, and for setup recommendations it stops checking altogether. Never present a dismissed recommendation as pending work. 'suggested' and 'approved' are the OPEN states: work that still needs attention. 'done' and 'dismissed' are closed. Filtering: - status, category, kind and priority each accept a COMMA-SEPARATED list (e.g. status='suggested,approved', category='content_gap,setup'). - status also accepts the special value 'open', which expands to every open state (suggested + approved). Prefer status='open' over spelling those two out — it stays correct if FirstAnswer adds another open state. It combines freely, e.g. status='open,done'. - start_date / end_date (YYYY-MM-DD) filter on when it was created. - order_by: '-created_at' (default, newest first), 'created_at', '-priority' (high first), '-completed_at' (most recently completed first). This tool is valuable for: - Answering "what should I do to improve my AI visibility?" (status='open', order_by='-priority') - Reviewing the outstanding backlog by priority - Auditing what was already completed in a period (status='done', order_by='-completed_at') The list omits the recommendation body. Call get_recommendation with an id to read the full payload (the instructions, drafted files, URLs, etc.).
list_responses
ChatGPTList individual AI responses across ALL monitored prompts of a brand, one row per response, with the URLs each response cited. Unlike get_monitored_prompt_responses this does NOT require a monitored_prompt_id, so it is the general-purpose data tool for anything computed over the whole response population. The unit of analysis is the RESPONSE. That matters: aggregated tools (list_cited_sources, list_cited_urls) report a count per domain/URL, so summing them double-counts a response that cited two of them and their overlap is not recoverable. Here each response appears once, with its own cited_urls, so any set logic the caller needs is exact. Counting instead of paginating `pagination.total_records is the number of DISTINCT responses matching the filters. When you only need a count, ask for one row and read the total instead of walking pages: list_responses(..., fields=["id"], per_page=1) -> pagination.total_records Any combination of filters works the same way, so a rate is two such calls (filtered over unfiltered). For example, responses citing at least one host from a list, out of all eligible responses: list_responses(..., cited_domain=["a.com", "b.com"], fields=["id"], per_page=1) list_responses(..., fields=["id"], per_page=1) Add brand_mentioned=true to either call to restrict to responses that mention the brand. Unfiltered, total_records reconciles with get_dashboard total_responses. Filters monitored_prompt_id, group_id, ai_model and cited_domain all accept multiple values. brand_mentioned restricts to responses that do/do not mention the brand. Unknown prompt or group ids return 404 instead of an empty list, so a mistyped filter is never mistaken for "no data". cited_domain is an EXACT hostname match, ORed across values: "example.com" does not match "www.example.com" or "uat.example.com", so a list of hosts stays literal and neighbouring subdomains are never swept in. List each host explicitly and confirm spelling with list_cited_sources — a host nobody cited simply matches nothing rather than erroring. Payload Use fields` to control what comes back — responses are large: - Default (fields omitted): id, monitored_prompt_id, ai_model, created_at, brand_mentioned, brand_position, brand_sentiment, cited_urls - Opt-in: prompt_text, groups, cited_domains, search_queries, response_text, response_markdown fields=["cited_urls"] gives the minimal response id -> citations base. Only ask for response_text / response_markdown when you need to read the answers. Field notes: - cited_urls: distinct NORMALIZED URLs (lowercased host, no fragment, no trailing slash, utm_source/srsltid stripped) in citation order. - brand_position / brand_sentiment come back as strings (API-wide decimal convention); cast before doing arithmetic. Dates are YYYY-MM-DD and are compared against the analysis run date, the same window every other tool uses.
remove_competitors_from_category
ChatGPTRemove monitored competitors from a competitor category. Only the association is removed: the competitors keep being tracked and the category itself is not deleted. WRITE ACTION — confirm with the user before calling. - competitor_category_id: from list_competitor_categories. - monitored_competitor_ids: 1 to 200 ids from list_monitored_competitors. Unknown ids reject the whole call. Returns removed_competitor_ids, not_in_category_ids and the category's full monitored_competitor_ids after the change.
remove_prompts_from_group
ChatGPTRemove monitored prompts from a prompt group. Only the association is removed: the prompts keep being monitored and keep their history, and the group itself is not deleted. WRITE ACTION — confirm with the user before calling. - group_id: from list_groups. - prompt_ids: 1 to 200 ids from list_monitored_prompts. Unknown ids reject the whole call. Returns removed_prompt_ids, not_in_group_ids and the group's prompts_count.
set_monitored_competitor_website
ChatGPTSet the website of a monitored competitor (add it when missing, or replace it). Nothing else about the competitor changes: its name, mentions and dashboard numbers stay as they are. Competitors cannot be renamed here. WRITE ACTION — confirm the website with the user before calling. - monitored_competitor_id: from list_monitored_competitors. - website: full URL, e.g. https://example.com. Returns {id, name, website}.
update_competitor_category
ChatGPTEdit the name, description and/or color of a competitor category. Omitted fields are kept as they are; its competitors are not changed (use add_competitors_to_category / remove_competitors_from_category). WRITE ACTION — confirm the change with the user before calling. - competitor_category_id: from list_competitor_categories. - name: up to 60 characters, unique within the brand (case-insensitive); a clash fails with competitor_category_already_exists. - description: up to 500 characters; an empty string clears it. Returns {id, name, description, color, created_at, monitored_competitor_ids}.
update_prompt_group
ChatGPTRename an existing prompt group. Its prompts are not changed (use add_prompts_to_group / remove_prompts_from_group for that). WRITE ACTION — confirm the new name with the user before calling. - group_id: from list_groups. - name: the new name, unique within the brand (case-insensitive); a clash with another group fails with group_already_exists and that group's id. Returns {id, name}.
get_brand
Claudeget_cited_source_related_prompts
Claudeget_cited_source_url_related_prompts
Claudeget_cited_source_urls
Claudeget_dashboard
Claudeget_monitored_competitor
Claudeget_monitored_prompt
Claudeget_monitored_prompt_citations
Claudeget_monitored_prompt_responses
Claudeget_monitored_prompt_search_queries
Claudeget_recommendation
Claudelist_brands
Claudelist_cited_sources
Claudelist_cited_urls
Claudelist_groups
Claudelist_monitored_competitors
Claudelist_monitored_prompts
Claudelist_recommendations
Claude