TrendsFast

Query the shared saved trend index without Product DNA. Maximum 1 credit(s); zero provider calls. Results rank explicit popularity or measured growth first, then US, UK, Canada and Australia, then Europe, alternating sources. data.brief groups the items into trending now, audience conversations, creative references, search demand and references, and names coverage gaps and next steps. Each item's evidence_class, uses and caution give what it supports and what it does not establish. Saves a result and debits credits when charged. The same request_id returns that saved result without a second charge; a new request_id is a new billable query.

Rank evidence against confirmed project context or explicit temporary niche. Without product: maximum 3 credits and no new source research. A product needs the confirmed context_id: it first collects the brand's own sources fresh, then ranks through that product's lens: customers (Customers) 10 credits, niche (Trending) 25 credits, competitors (Competitors) 30 credits, formats (Formats) 20 credits, search (Search) 35 credits. A product on a typed niche is refused with PRODUCT_REQUIRES_CONTEXT and charges nothing; a typed niche is read without a product. Formats also reads the top TikTok and YouTube videos' transcripts, so each item's hook is its opening line. Search also asks ChatGPT, Claude, Gemini and Perplexity the category's buyer question (no brand is sent) and returns data.ai_answers: who each answer names, whether that includes the brand or its competitors, and the domains cited. A Customers or Trending call, or a read without product, also returns data.communities when the brand's confirmed communities have been read: each one's rules on self-promotion, links and AI-written posts (citing rule numbers), the formats it favours, what wins there over all time (winning_format, hook_pattern and avoid, read by Claude from its top posts), the moderators' pinned posts (pinned), and a play for taking part without breaking them. A Trending call may also give each item a topic_history: where its topic is in its life across days and channels (state, since, first_seen, peak_day, channels, 7-day velocity, as_of), or null until the topic has a lifecycle state; it is the only basis for saying when a topic started rising. new_research says whether the call collected or asked. CORE means a specific topic, problem, keyword or exact competitor-account match; sharing only a broad family such as AI is ADJACENT. A temporary niche may add communities, competitor_handles, competitor_domains and keywords for exact matching. data.brief and each item's evidence_class, uses and caution d

Answer one question across global, core and adjacent scope and every source channel in a single call. Maximum 5 credits; zero provider calls. Global scope needs no Product DNA. A channel with no permitted observation is reported UNAVAILABLE rather than empty. Saves a result and may debit credits. No new source research.

Answer a bounded question using saved intelligence. Maximum 5 credits. Questions about competitors, customer complaints, feature requests, hooks and formats, or search keywords each need their own evidence: answer_status says ANSWERED, PARTIAL, INSUFFICIENT_EVIDENCE, NEEDS_CONTEXT or NOT_COLLECTED, findings cite each item with its kind, and missing_inputs and next_steps say what would complete the answer. An answer with no findings costs nothing. Saves a result and may debit credits. No provider or model calls, invented growth or hidden fresh research.

Read a saved result, page it on next_cursor, or recover an exact accepted request ID. No credit debit or research. Source permissions, retention and context are rechecked.

Read the connected project name, ID and current confirmed niche context, or an explicit missing-context state. Returns the context_id that niche and Ask queries accept. No credit debit, context mutation or new research.

無需輸入參數。

Read the shared REST/MCP credit balance, monthly/prepaid split, reservations and pause state without spending credits.

無需輸入參數。