seoClarity ArcAI

Get brand profile. By default returns all profiles including competitors. Set includeCompetitors to false to see only the own brand profile (isOwnBrand:true). Response fields: - id: Brand profile ID - profileName: Brand profile name - isOwnBrand: Whether this is the own brand profile - competitorList: List of tracked competitors

Ask a question about seoClarity's ArcAI products and features. This tool queries the seoClarity ArcAI knowledge base to answer questions about: - ArcAI product features and capabilities - How to use ArcAI Visibility, Content Optimizer, Indexer, etc. - Best practices and workflows - General seoClarity platform questions Response: - Returns a detailed answer from the knowledge base - May include relevant citations and sources

Get list of prompts tracked in ArcAI Visibility. Can filter by topic or tag. Response fields: - id: Prompt ID - prompt: The prompt text being tracked - topics: Array of topics this prompt belongs to - tags: Array of tags assigned to this prompt Results are PAGINATED and limit caps at 100 - a domain can track thousands of prompts. A full page usually means more pages exist: keep paging with offset until a page comes back short or empty before concluding a prompt or tag is absent. To find prompts carrying a given tag, do NOT scan pages by eye - resolve the tag ID with 'arcai_tags' and pass it in 'tags' so the server does the filtering.

Get raw AI engine responses, sentiment, and citations for prompts within a date range. One row per prompt x engine — the same prompt appears once per engine, each with its own response. Fields per row (limit/offset count rows, not prompts): - prompt: The prompt text - engineId/engineName: Which AI engine responded - response: The engine's full text answer - promptResponseType: 'search_augmented' (grounded in cited sources) or 'model_knowledge' (from training data) - brands: Brand names mentioned, as plain strings. Extraction is noisy — it can include non-brands such as '12-month standard warranty' - citations: Array of cited-source objects, each { url, domain_type, domain_category }. domain_type is 'Own Brand Profile', 'Competitor Brand Profile', or 'Neither' (useful for share-of-citation analysis); domain_category is a list of site categories, e.g. ['Media & Publications'], and may be empty - products: Products mentioned, as { name, brand, url }. This is a thin projection of response_features.products with 'attributes' stripped — same elements, same order. Use response_features.products for the enriched form - queryFanouts: The search queries the engine fanned out to before answering, as plain strings. Often empty, and it does NOT track promptResponseType in either direction: 'model_knowledge' rows can carry fanouts, 'search_augmented' rows can have none - inlineHyperlinks: Hyperlinks embedded in the response, as [{ text, link }] pairs. Anchor text is NOT unique — the same text may appear with different links - date: Date of the response - languageId/languageName/languageCode: Language of the response - topics/tags: Associated topics and tags, each { id, name } response_features: SERP-style features extracted from the response. Always present, and all four keys are always present as arrays — emptiness is [], never omission: products, ads, maps, jobs. Every element is { name, brand, url, attributes }, and every value inside attributes is a string. Only ChatGPT (engineId 1) populates these; other engines return the four empty arrays. promptResponseType does NOT predict whether features are present. 'brand' is empty on ads, maps and jobs — the real name is in attributes (brandName for ads, company for jobs); only products populates it. The keys listed per type below are the full set; an individual key may be ABSENT — an unavailable value is omitted from attributes entirely, never returned as null or an empty string: - products.attributes: img, price, rating (0-5), review (count), carouselProduct (STRINGIFIED JSON array of { rank, merchantName, merchantUrl, merchantType, price, desc } — parse before use; '[]' when empty) - ads.attributes: img, brandName, domainUrl, faviconUrl. faviconUrl may be '[omitted: inline base64 favicon]' — inline base64 favicons are dropped before returning, because one observed row was 90% favicon bytes. Ordinary URLs pass through - maps.attributes: img, latitude, longitude, address, rating (0-5), review (count), categories (STRINGIFIED JSON array of category names — parse before use), phone, location (the surrounding region, not the street address) - jobs.attributes: roleSummary, company (the employer), source (job board, e.g. 'indeed'), location, salary, employmentType. salary and employmentType are human-readable display strings ('$90,000 - $95,000 a year', 'Full-time'), not structured values Envelope: the response also carries meta { limit, offset, hasNext } — page with hasNext. limit defaults to 10 and caps at 100. The date range cannot exceed 3 months — a wider range is rejected with HTTP 400, code API-3000064.

Get prompt-level sentiment data with per-engine sentiment values. PREREQUISITE: Requires brandProfileId. Call 'arcai_brand_profile' first to get it. Response fields per prompt: - prompt: The prompt text - searchVolume: Search volume - chatgpt_sentiment, aimode_sentiment, aioverview_sentiment: Sentiment per engine (positive/neutral/negative/mixed/-)

Get aggregate sentiment percentages (positive, neutral, negative, mixed) trended over a date range. PREREQUISITE: Requires brandProfileId. Call 'arcai_brand_profile' first to get it. Response fields per data point: - date: Data point date - sentiment_positive_percent, sentiment_neutral_percent, sentiment_negative_percent, sentiment_mixed_percent - sentimentTotalCnt: Total prompt count

Get list of tags defined in ArcAI Visibility for this domain. Use this to find the numeric tag IDs that the 'tags' parameters and filters on the other tools require - filtering by tag is far cheaper and more reliable than listing prompts and scanning their tags arrays. Results are PAGINATED and limit caps at 100. A full page usually means more pages exist - keep paging with offset until a page comes back short or empty before concluding a tag does not exist.

Get list of topics tracked in ArcAI Visibility for this domain. Response fields: - id: Topic ID - name: Topic name - promptsCount: Number of prompts in this topic - lastUpdated: Last update timestamp Results are PAGINATED and limit caps at 100. A full page usually means more pages exist - keep paging with offset until a page comes back short or empty before concluding a topic does not exist.

Get prompt-level visibility and citation data with per-engine metrics. PREREQUISITE: Requires brandProfileId. Call 'arcai_brand_profile' first to get it. IMPORTANT: Use startDate and endDate parameters (NOT queryDate). Response fields per prompt: - prompt, date, languageCode, language_id, searchVolume - topics: Array of { id, name } - mentionPercent, citationPercent: Overall mention/citation percentages Per-engine fields use the engine name as a prefix; which engines appear depends on the domain's tracked engines. All possible prefixes: chatgpt, aimode, aioverview, claude, gemini, perplexity, copilot. For each engine <e>: - <e>MentionPercent, <e>CitationPercent: Mention/citation percentages for that engine - <e>Content: The engine's full text answer - <e>Citations: Array of cited-source objects, each { url, domain_type, domain_category } (domain_type = 'Own Brand Profile' | 'Competitor Brand Profile' | 'Neither'; domain_category is a list of site categories, e.g. ['Digital Services & Tools'], and may be empty) - <e>PromptResponseType: 'search_augmented' or 'model_knowledge' - <e>PromptType: 'branded' or 'non_branded'

Get aggregate visibility and citation metrics trended over a date range across AI engines. PREREQUISITE: Requires brandProfileId. Call 'arcai_brand_profile' first to get it. Response fields per data point: - date: Data point date - allEngineMentionPercent/Count, allEngineCitationPercent/Count: Overall metrics - chatgptMentionPercent/Count, chatgptCitationPercent/Count: ChatGPT-specific - aimodeMentionPercent/Count, aimodeCitationPercent/Count: AI Mode-specific - aioverviewMentionPercent/Count, aioverviewCitationPercent/Count: AI Overview-specific