HypeAuditor

Social listening and Media Monitor of the connected HypeAuditor account: which listeners run, the creators and posts they found mentioning the brand, Media Monitor feeds and their latest content. Start or pause an existing listener. Free — the user's own data. Creating listeners and feeds stays in the HypeAuditor app.

hypeauditor-listening

The connected HypeAuditor account's media plans: list plans, read a plan's creators and forecast metrics (reach, CPM, budget), create or rename a plan, add or remove creator accounts. Free — the user's own data. Accounts are addressed by platform + social_id (hypeauditor-get-account-social-ids).

hypeauditor-media-plans

The connected HypeAuditor account: subscription, what is left of each plan limit (report unlocks, Market Analysis credits, email unlocks, exports, AI credits), and team members. Free. Use it when the user asks what their HypeAuditor plan includes or how much of a limit remains. HypeAgent credits are a separate balance — see hypeagent-account.

hypeauditor-my-account

The connected HypeAuditor account's My Network (creator CRM): lists, the creators in them, search, and adding creators to the network or a list. Free — this is the user's own data. Use it to answer "is @x already in my network?", to build shortlists, and to get the influencer_id that hypeauditor-campaigns.add_influencers needs.

hypeauditor-my-network

The user's HypeAuditor influencer reports. USE action "full" WHENEVER THE USER ASKS FOR A FULL / COMPLETE HypeAuditor REPORT on one creator ("give me the full report for @x", "unlock @x", "open the HypeAuditor report") — it is paid from the HypeAuditor report balance (one report; the plan's analytical reports and purchased credits are ONE pool, shown as "advanced report tokens" in hypeauditor-my-account; free if already unlocked), NEVER from HypeAgent credits, and the creator stays unlocked in the user's HypeAuditor library. It answers confirmation_required before spending a token; ask the user, then repeat with confirm: true. Also: list unlocked reports, read one for free, unlock several at once, unlock creator emails. For a quick analysis or a specific question about an account that is NOT unlocked, hypeauditor-analyze-accounts / hypeauditor-get-account-insights (HypeAgent credits) are the tools instead.

hypeauditor-my-reports

List every HypeAuditor discovery filter that is NOT already a parameter of hypeauditor-search-influencers, with its exact name, type, allowed values and range. FREE — costs no credits. Call this before using that tool's `advanced` object, and whenever a user asks for something the visible parameters do not cover (creator verification or account age, audience ethnicity or income, category and thematic ids, average views, posting cadence, sponsored-post counts, estimated post price, brand mentions, bio and caption text search, brand safety). Pass `platform` to also get that platform's numeric category, thematic and audience-interest ids — the vocabularies category_include/exclude, blogger_thematics_include/exclude and audience_thematics take. Pass scope "all" to also list the parameters already in the search tool's own schema.

hypeauditor-search-filters

Search/discover influencers in the HypeAuditor database (Instagram, YouTube, TikTok). ⚠️ ONLY set parameters the user EXPLICITLY asked for. Every filter you add narrows the results — a filter the user never mentioned over-constrains the search and returns ZERO results. When in doubt, leave a parameter out. Example: "find instagram crypto creators 10-50k followers" → set ONLY platform, niche, followers_count_from, followers_count_to. Nothing else. The reverse is equally wrong: when the user DOES state a constraint, set it, and never tell them a filter does not exist without checking hypeauditor-search-filters first. **CAPABILITIES (use only when relevant):** - niche + keywords: the main relevance signal (e.g. niche "crypto", keywords ["bitcoin","defi"]). `niche` is one semantic query: space-separated words blend into a single topic, so separate distinct topics with commas ("forex trading, stock investing") — never a space-separated keyword list. Write `niche` in the language the creators post in — it is matched semantically, so "beleza" finds Brazilian beauty creators that "beauty" ranks lower. - follower range, engagement rate, audience quality score - creator: blogger_gender, blogger_age_from/to, blogger_languages, account_country_code / account_countries_include (based in X) - audience: audience_gender (+prc), audience_age_groups (+prc), audience_geo (audience located in country X — COUNTRIES only, never a city) - account_owner_type "creator": drops brands, shops and salons — set it when the user wants people, not businesses - look-alike: similar_to_accounts ("more creators like @x") - exclusions: social_id_exclude / username_exclude — use when iterating so the same creators are not returned again - last_active_within_days + blogger_has_contacts: for reachable, active creators before outreach - growth: minimum 30-day follower growth % - paging: `from` — repeat the search with from = 0, limit, 2×limit … for the next page (each page is a new search) - contacts: include_emails — ONLY when the user asks for emails; priced per creator returned with emails - anything else: ~45 further filters (verification, account age, audience ethnicity/income, category ids, average views, posting cadence, sponsored posts, post price, brand mentions, bio/caption text search, brand safety) are listed by hypeauditor-search-filters (FREE) and passed through `advanced`. **BEFORE SAYING SOMETHING IS IMPOSSIBLE:** call hypeauditor-search-filters. It costs nothing and it is the full list. A capability missing from this schema is usually in `advanced`, not missing from the product. **TIP:** Start broad. If too many results, add ONE more filter at a time. Do not pre-emptively stack filters.

hypeauditor-search-influencers

Search and retrieve social media content from Instagram, YouTube, and TikTok with advanced filtering. NOT for a link to one post: the index matches captions, never URLs or shortcodes, so that search finds nothing and still costs credits — use hypeauditor-get-post for a post link. **WHEN TO USE:** - User asks "show me posts mentioning @brand" or "find mentions of [brand]" - User asks for "trending content about [topic]" or "viral posts with #hashtag" - User wants to find "sponsored posts" or "paid partnerships" for a brand/creator - User asks "who is talking about [brand/product]" - User wants to track "competitor mentions" or "industry buzz" - User needs to see "content with [keyword]" or "posts using #hashtag" **USE CASES:** 1. **Brand Mentions**: Find content mentioning a brand (via @mentions of its username, and the brand name as keyword and hashtag) 2. **Hashtag Tracking**: Search content using specific hashtags 3. **Trending Discovery**: Find high-engagement content in a niche 4. **Sponsored Content**: Filter for paid/sponsored posts only 5. **Competitor Research**: Monitor where competitors are being mentioned **DO NOT USE when:** - User wants ONE known account's own recent posts ("show me @username's recent posts") → use hypeauditor-analyze-accounts; its recent_content section has the posts with permalinks and thumbnails - User wants account analytics/metrics → use hypeauditor-analyze-accounts or hypeauditor-get-account-insights - User wants to compare influencers → use hypeauditor-compare-accounts - User wants to discover NEW influencers → use hypeauditor-search-influencers **PERIOD:** - Pass `date_from` / `date_to` (YYYY-MM-DD, both inclusive) for a period: "in March 2025" is date_from 2025-03-01, date_to 2025-03-31. `search_depth_days` is shorthand for "the last N days". With neither, the last 90 days are searched. - The index keeps posts from the 1st of the month two years ago. An earlier `date_from` is searched from there, and the response `note` says so — tell the user the period is shorter than asked. - One search pages through its first 10 000 rows at most. For more, split the period (e.g. one month per search) and page each. **READING RESULTS:** - `total_found` is how many posts the INDEX holds for the strongest filter the search used; `returned` is how many are in this response. Never report `returned` as the number of mentions when `total_found` is larger — say how many exist and how many you are showing. - To collect more, repeat the same search with `offset` set to the `next_offset` of the previous response. When `next_offset` is absent there is nothing left to read, however much larger `total_found` is — `total_found` counts index rows, and a `brand_mentions` search drops the ones that do not actually name the brand. `returned` can also be smaller than `limit` on a page that still has a `next_offset`. - `total_by_filter` breaks that down per filter (`mentions`, `hashtags`, `keywords`). The `keywords` figure is a TEXT match and is routinely inflated — a short brand name matches longer words — so never quote it as a mention count. - On `brand_mentions` every post carries `matched_by`: `mention` (the post tags the account — the reliable one), `hashtag` (carries its tag), `name` (the caption only says the words). Say which when it matters; do not call a `name` match a mention. - `searched_account` says which account the mention filter actually ran on, because the index answers only to an exact username. If `resolved` is false, a zero result may mean the handle is wrong rather than that nobody mentions the brand — say so instead of reporting zero mentions. - `basic.post_date` can be the date HypeAuditor INDEXED the post, not when it was published (backfilled rows share one timestamp). Prefer `basic.created_at_from_id` (decoded from the post id; TikTok and Instagram) for when a post was actually published. - `engagement_rate_percent` is a percent (5.13 means 5.13%). `is_sponsored` is absent when the platform does not report it — that is unknown, not "not sponsored". - The index covers creators HypeAuditor tracks, not all of the platform — a small `total_found` means few tracked creators posted about it, not that nobody did.

hypeauditor-content-search