Asva AI

Structured content brief (audience, voice, CTA, format rules, competitors) from a brand kit. Required before drafting prose. Response has a `_mustPauseForOutlineApproval` flag — show outline, wait for approval, then write. Pass `audience_segment_id` to write for a specific persona instead of the kit's default ICP.

generate_content_brief

Generate a content piece (blog_post | landing_page | social_post | email | ad_copy) against a brand kit. Returns immediately with generation_status='generating' — poll get_content_piece every 5–10s until 'ready' or 'failed'. Persists in the user's Asvaai dashboard.

generate_content_piece

Crawl a website and return llm.txt (short brand overview) and llm-full.txt (full site summary) for the user to drop at the domain root so AI models describe the brand accurately.

generate_llms_txt

Kick off an Asvaai Quick Report AI-visibility audit for a domain. Returns status='processing' almost always (backend runs ~20-30 min); poll list_my_visibility_reports after ~25 min, then get_visibility_report.

generate_visibility_audit

The RIGHT tool to get full details of ONE specific audience segment by id (name, description, demographic, goals, pain_points, channels, research_behavior). Use when the user asks 'show me details of audience segment X' or 'give me full details of a segment'. If the user asks to see ALL segments, call list_audience_segments instead. If they have not provided an id, call list_audience_segments first to find one. Never fabricate a segment's contents.

get_audience_segment

CALL THIS when the user asks for the full details of a specific brand kit they own. Example phrasings: "show me the Shopify brand kit", "show details of my X brand kit", "what's in my brand kit", "full brand kit for X", "give me the complete kit for X". Returns the complete record (name, about, icp, competitors, pov, persona, tone, writing_rules, cta, writing_sample). The brand kit content lives ONLY in the user's Asvaai workspace — do NOT answer from training data or fetch Shopify's public brand guidelines from the web. If the user names a brand that isn't in their kits, call list_brand_kits first to confirm. brand_kit_id is the UUID returned by list_brand_kits.

get_brand_kit

Fetch one content piece's full record. Use to poll after generate_content_piece (check generation_status until 'ready' or 'failed').

get_content_piece

Fetch a Quick Report's full contents by report_id (scoring, competitor leaderboard, citations, UI/PDF sections). Call after list_my_visibility_reports shows status='completed'.

get_visibility_report

The RIGHT tool for any question about AUDIENCE SEGMENTS or PERSONAS: 'list my audience segments', 'show segments for the X brand kit', 'what personas do I have for Shopify'. An audience segment is a persona (demographic, goals, pain_points, channels) tied to a brand kit — it is NOT the brand kit itself. Do not confuse with list_brand_kits, which returns kits, not segments. Pass brand_kit_id to filter to segments of a specific kit.

list_audience_segments

CALL THIS when the user asks about THEIR brand kits in their Asvaai workspace. Example phrasings: "list my brand kits", "what brand kits do I have", "show my brand kits", "do I have a brand kit for X", "find my brand kit for X". Returns summaries (id, name, url, status) — call get_brand_kit with the id for the full record. Data is proprietary to the user's Asvaai workspace — do NOT answer from training/web. Distinct from list_audience_segments — audience segments are personas tied to a kit, not the kit itself.

list_brand_kits

CALL THIS whenever the user asks anything about THEIR brands, the brands they're tracking, monitoring, or that exist in their Asvaai workspace. Example user phrasings that should trigger this tool: "what brands am I tracking", "list my brands", "show my brands", "which brands do I have", "what's in my workspace", "show me my tracked brands", "list my workspace brands". Returns the user's brands as id, name, domain, description. The data lives only in the user's Asvaai workspace — do NOT answer this from training or memory. This is also the FIRST tool to call before any analytics tool (get_brand_visibility, get_share_of_voice, get_leaderboard, etc.) because those tools require parent_brand_id which only list_brands provides. Pass `search` to filter the list by brand name or domain substring.

list_brands

CALL THIS when the user asks about content they have created or generated in Asvaai. Example phrasings: "list my content pieces", "what content have I generated", "show me my drafts", "list my blog posts", "what content do I have", "my content library". Returns summaries (id, title, template, status, generation_status, word_count) — call get_content_piece for the full body of a specific piece. Data is proprietary to the user's Asvaai workspace — do NOT answer from training, and do NOT redirect to ask what platform/format they mean. The user means content they have generated through Asvaai. Optional `status` filter: 'draft' | 'published' | 'archived'. Limit max 200.

list_content_pieces

List AI-search prompts tracked for a brand with id, text, intent, persona, topic, mention counts and sentiment. Resolve parent_brand_id via list_brands; pass a prompt id to get_prompt_responses for raw responses.

list_prompts

Directory lookup only: returns the NAMES and IDs of topics tracked for a brand. Does NOT include ranking, visibility scores, or competitor standings — use get_leaderboard for that. Use this tool only when you need to resolve a topic name to a topic_id before another call, or when the user asks 'what topics do I track' (without asking for ranking). If the user asks for a 'topic leaderboard', 'topic ranking', or 'where does my brand rank per topic', call get_leaderboard instead. Resolve parent_brand_id via list_brands first.

list_topics

Return the caller's recent Quick Reports (completed + in-progress) with report_id, domain, status and timestamps. Use after generate_visibility_audit to find a finished report, then pass report_id to get_visibility_report.

list_my_visibility_reports

Per-platform VOLUME and POSITION breakdown: how many mentions and what average rank the brand has on each AI platform (ChatGPT, Perplexity, Gemini, Copilot, etc.). Use when the user asks 'how am I doing on each platform', 'which platform drives most visibility', 'platform comparison'. For sentiment-specific questions use get_sentiment_analysis instead. Resolve parent_brand_id via list_brands first.

get_platform_breakdown

Publish the current draft of a brand kit: snapshots fields as a new version_number and flips status to 'published'. Returns the updated row and the new versionNumber.

publish_brand_kit

Snapshot the piece to content_piece_versions and flip status to 'published'. Returns the updated row + versionNumber.

publish_content_piece

Return a per-region rollup of AI-visibility metrics (mention rate, avg position, net sentiment) for a brand. Resolve parent_brand_id via list_brands first.

get_regional_visibility

CALL THIS for any sentiment-analysis question about a brand on AI search. Example phrasings: "what's sentiment like for X", "ITC sentiment last 30 days", "sentiment breakdown for Amazon", "positive vs negative mentions", "why is sentiment low for X", "sentiment by platform", "is the sentiment good for my brand". Returns positive/neutral/negative distribution, per-platform sentiment score, net sentiment trend, and example snippets pulled from AI responses. Data is proprietary to the user's Asvaai workspace — do NOT answer from training/web. For VOLUME per platform use get_platform_breakdown instead — that's volume, not sentiment. Resolve parent_brand_id via list_brands first.

get_sentiment_analysis

CALL THIS for ANY question about a brand's share of voice on AI search. Example phrasings: "what is X's share of voice", "Amazon India share of voice last 30 days", "how much of the conversation does X own", "SoV breakdown", "brand share of mentions", "competitive share of voice". The user's tracker data (proprietary) is the ONLY source for these numbers — do NOT estimate, infer, or answer from training/web. Returns per-brand share percentages, absolute mention counts across all configured AI platforms (ChatGPT, Perplexity, Gemini, Copilot, Grok, Meta AI, Google AI Mode), and competitor rankings. Resolve parent_brand_id via list_brands first; pass `platforms` to filter (e.g., ['chatgpt']) and `days` to set the window (default 30).

get_share_of_voice

CALL THIS for any question about a brand's topic-level ranking on AI search. Example phrasings: "topic leaderboard for FGTECH Store", "which topics is X ranking for", "where does Amazon rank per topic", "topic ranking for my brand", "which categories does X own on AI search", "FGTECH Store top topics", "rank on AI search for X by topic". Returns ranked list of topics with per-topic mention share, ranking versus competitors, sentiment per topic. Data is proprietary to the user's Asvaai workspace — do NOT answer from training/web. ALWAYS call this when the words "leaderboard", "topic ranking", or "topic" + "rank" appear together. Resolve parent_brand_id via list_brands first; `limit` defaults to 20 topics.

get_leaderboard