目录 / Unyly Gateway
MCP
鉴权未知
未评级
已上架
Unyly Gateway
One connector for the entire MCP catalog — 15,000+ servers plus your team's private MCPs — callable from Claude, ChatGPT, Cursor, VS Code and any MCP client. Search, list and call any server through a single OAuth endpoint. No per-server install, no JSON config.
该来源不提供完整文件导出（国内平台多为平台内托管），仅存元数据与原链
接入信息
- 传输形态
- http
- 鉴权方式
- 鉴权未知
- 端点
https://gateway--unyly.run.tools
鉴权方式未标注，请核对官方文档后再接入——不要直接使用以下片段
{
"mcpServers": {
"Unyly Gateway": {
"url": "https://gateway--unyly.run.tools"
}
}
}
能力清单
| 工具 | 说明 |
|---|---|
| search_mcps | Search the Unyly MCP catalog (47,000+ servers) by task, name or category. Returns matching MCPs: slug, name, what they do, and reachability ([remote] = callable through the gateway; [hosted] = local/stdio, may not be runnable yet). Use this first when you need a capability you don't already have as a `<slug>__<tool>` tool, then call list_mcp_tools on a [remote] slug. |
| list_mcp_tools | Unyly gateway — list the tools of a catalog MCP by slug (from search_mcps), so you can see what it can do before calling use_mcp_tool. If it replies that the server is hosted/stdio and not enabled, that MCP can’t be called through the gateway yet — pick a [remote] one instead. |
| use_mcp_tool | Unyly gateway — call a tool on any of the 47,000+ MCP servers in the Unyly catalog. Provide mcp_slug, tool, and arguments (object). "Payment required" means the user’s team/wallet balance is empty — tell them to top up at unyly.org, it is not an error on your side. "hosted/stdio not enabled" means that server can’t run through the gateway yet — use a [remote] one. |
| nectarin-intelligence-worker__ru_benchmarks | [NECTARIN Intelligence] RU/CIS advertising benchmarks. Returns CPM/CTR/CPA/VTR ranges (p25/p50/p75) and percentile context for a category × KPI, optionally narrowed to one platform (VK Ads, Yandex Direct, Telegram Ads, OLV, Avito). Categories incl. realty, finance, auto, retail, fmcg, pharma, ecom, edtech. Includes data provenance. Mock aggregated data. |
| nectarin-intelligence-worker__supplier_quality | [NECTARIN Intelligence] Inventory / supplier quality index for RU/CIS. Filter by format and/or platform (optionally a category) to get a 0-100 quality score, fraud-risk rating, viewability/human-traffic signals, and a list of recommended (clean) formats plus suppliers to avoid. Mock data. |
| nectarin-intelligence-worker__media_plan | [NECTARIN Intelligence] Build a RU/CIS media plan. Distributes a RUB budget across VK Ads / Yandex Direct / Telegram Ads / OLV based on the marketing goal, then computes a real forecast from mock benchmarks (impressions = spend / CPM * 1000, clicks = impressions * CTR, conversions = spend / CPA, plus estimated reach and blended CPA). Returns channel split %, flighting, forecast totals, per-channel detail, narrative rationale, and compliance flags. |
| nectarin-intelligence-worker__category_playbook | [NECTARIN Intelligence] Category go-to-market playbook for RU/CIS: communication territories, do's & don'ts, seasonal hooks, and compliance notes. Regulated categories (pharma, finance) get a STOP-GATE flag requiring legal sign-off before launch. Mock data, not legal advice. |
| nectarin-intelligence-worker__audience_insights | [NECTARIN Intelligence] Audience insights for a RU/CIS category: key segments (with relative size and notes), Jobs-To-Be-Done, and media affinities (which channels each audience leans into). Optional geo refinement. Mock data. |
| nectarin-intelligence-worker__competitor_scan | [NECTARIN Intelligence] Competitive landscape scan for RU/CIS. Provide a brand and/or category to get a list of likely competitors with estimated media activity, primary channels, and the communication territory each tends to own. Synthetic estimates — wire to real SOV monitoring (Mediascope, Brand Analytics) in production. |
| nectarin-intelligence-worker__geo_aeo_audit | [NECTARIN Intelligence] GEO / AEO (Generative & Answer Engine Optimization) audit for a brand: estimated visibility score inside AI answer engines and RU search — Yandex (neuro), GigaChat (Sber), YandexGPT/Alice, ChatGPT — plus concrete recommendations to improve how models describe and cite the brand. Synthetic scores. |
| nectarin-intelligence-worker__creative_brief | [NECTARIN Intelligence] Generate a creative brief for a product, audience and channel: objective, single-minded proposition, tone, mandatories (incl. RU ad-labeling), channel craft notes, and 3 distinct concept territories. Narrative comes from the (stubbed) LLM grounded on structured inputs. |
| nectarin-intelligence-worker__report_explain | [NECTARIN Intelligence] Explain a campaign metrics report in plain language. Accepts a JSON string of metrics (e.g. {"cpm":300,"ctr":0.4,"cpa":1800,"vtr":55}), returns a plain-language summary, detected anomalies (heuristic), and 3 prioritized recommendations. |
| nectarin-intelligence-worker__budget_optimizer | [NECTARIN Intelligence] Optimize a RUB media budget across VK Ads / Yandex Direct / Telegram Ads / OLV to MAXIMIZE conversions (not just follow a goal preset). Uses real optimization: conversions/RUB = 1/CPA, then water-fills the lowest-CPA channels first up to a per-channel cap (default 45%). Returns the optimal allocation, projected conversions & blended CPA, and the uplift vs. the goal-preset split. Mock benchmarks. |
| nectarin-intelligence-worker__strategy_orchestrate | [NECTARIN Intelligence] FLAGSHIP end-to-end orchestration. In ONE call NECTARIN fans out to all of its workers and returns a complete go-to-market strategy: RU/CIS CPA benchmarks, audience segments & JTBD, competitor landscape, goal-based channel split WITH a forecast (impressions/reach/conversions/blended CPA), a conversion-maximizing optimized split, a lead creative concept, compliance gate, a quick ROI framing, and an executive summary. The narrative uses a real LLM when LLM_API_KEY is set (Anthropic/OpenAI), otherwise a deterministic stub. Mock/synthetic data; not legal advice. |
| nectarin-intelligence-worker__roi_calculator | [NECTARIN Intelligence] SELL value. Projects the ROI of moving a media budget onto NECTARIN. Inputs: monthly_budget (RUB), optional current_cpa, and category. Output: current vs. projected CPA, extra monthly conversions, and estimated annual value — all derived from the same mock RU/CIS benchmarks (CPA p25/p50) the Intelligence tools use, with the method shown so the number is auditable. Synthetic/illustrative, not a guarantee. |
| nectarin-intelligence-worker__lead_qualify | [NECTARIN Intelligence] Qualify a marketer as a NECTARIN lead. Inputs: company, monthly_budget (RUB), industry, goal. Output: a 0-100 fit score, the recommended NECTARIN engagement tier (self-serve / managed / enterprise retainer) via budget thresholds + regulated-category and goal signals, and a plain rationale. Deterministic funnel logic — no external scoring service. |
| nectarin-intelligence-worker__request_nectarin_proposal | [NECTARIN Intelligence] Capture a structured RFP/brief for a NECTARIN engagement and return it for review along with a clearly-stubbed submission reference and next steps. Inputs: brief fields (company, industry, monthly_budget, goal, timeline, notes) + contact (name, email) provided BY THE USER. IMPORTANT: this does NOT send anything anywhere — no CRM write, no webhook, no email. It only echoes a structured brief + a local reference id. Wire the real CRM/webhook where marked in code. Do not include sensitive/special-category PII. |
| nectarin-intelligence-worker__book_consultation | [NECTARIN Intelligence] Return a scheduling CTA for a NECTARIN consultation: a booking URL (from env NECTARIN_BOOKING_URL, placeholder by default), the topic, an optional preferred_time echoed back, and a short 'what to prepare' checklist. No calendar API is called — this is a CTA generator, not a booking write. |
| nectarin-intelligence-worker__automation_recipe | [NECTARIN Intelligence] Describe a concrete multi-agent automation workflow NECTARIN can run for the client. Input: task (weekly_reporting | creative_variants | tender_deck | competitor_monitoring). Output: ordered steps, which INTERNAL NECTARIN tools each step calls, the cadence, the deliverable, and estimated time saved. This frames NECTARIN as managed automation, not just advice. |
| nectarin-intelligence-worker__value_forecast | [NECTARIN Intelligence] Three-scenario value projection (conservative / base / ambitious) of what NECTARIN + AI can deliver. Inputs: brand, budget (monthly RUB), horizon_months. Output: per-scenario reach, efficiency (CPA) and cumulative savings vs. status quo, with assumptions stated. Anchored to mock category-neutral benchmarks; deterministic. Illustrative, not a guarantee. |
| nectarin-intelligence-worker__compliance_check | [NECTARIN Intelligence] RU advertising-law compliance review of ad copy. Paste the creative text (+ optional category/platform) and get: a 0-100 compliance score, a list of flagged risks with severity, the relevant ФЗ-38 «О рекламе» article, and a concrete fix — covering superlatives/ФАС risk, comparative claims, finance (ПСК, guaranteed returns — ст. 28), pharma (mandatory warning — ст. 24), alcohol/tobacco/gambling hard-blocks, and ОРД/ЕРИР marking. When an LLM key is configured it also returns extra nuance and a compliant rewrite. Decision-support, NOT legal advice. |
| nectarin-intelligence-worker__ab_test_planner | [NECTARIN Intelligence] Plan an A/B test with real statistics (two-proportion z-test power analysis). Inputs: baselineRatePct (current conversion %), mdeRelPct (minimum relative uplift to detect, e.g. 10 = +10%), dailyVisitorsPerVariant, optional variants (default 2), powerPct (default 80), alphaPct (default 5, two-sided). Returns the required sample size per variant, total, estimated test duration in days, the detectable absolute lift, and guardrails (min runtime, multiple-comparison note). Deterministic — uses the inverse-normal (Acklam) for exact z-scores. |
| nectarin-intelligence-worker__unit_economics | [NECTARIN Intelligence] Marketing unit economics & health check. Inputs: aov (avg order value or ARPU per purchase, RUB), grossMarginPct, and EITHER cac directly OR (monthlySpend + newCustomers) to derive it; plus repeat behaviour as purchasesPerYear and lifespanYears (or churnRatePct). Returns gross-margin LTV, LTV:CAC, payback period (months), ROAS, contribution per customer, a health verdict (LTV:CAC ≥3 healthy, payback <12mo good), and concrete levers. Deterministic; illustrative, not a guarantee. |
| nectarin-intelligence-worker__funnel_model | [NECTARIN Intelligence] Model the FULL marketing funnel for a budget, end to end: impressions → reach → clicks → leads → qualified → sales → revenue, with conservative/base/optimistic (P10/P50/P90-style) scenarios derived from the benchmark spread (p75/p50/p25 CPM·CTR). Reports stage counts, drop-off at each step, CAC, ROAS and revenue when an AOV is given. Identifies the biggest leak. Deterministic; illustrative, not a guarantee. OPTIONAL: set `useTenantData:true` to derive real click→lead / lead→sale conversion rates (and AOV) from the tenant's funnel counts in KV (mode=kv only; strictly opt-in — absent/mock ⇒ unchanged category defaults). |
| nectarin-intelligence-worker__seasonality_forecast | [NECTARIN Intelligence] When to spend. Returns a 12-month demand/competition index for a RU/CIS category (mean ≈ 1.0), the peak and trough months, a recommended budget weighting across months, and a flighting recommendation (lean in before peaks, protect efficiency in troughs). Optionally splits a provided annual budget by month. Deterministic. |
| nectarin-intelligence-worker__creative_score | [NECTARIN Intelligence] Score an ad creative (headline + body, optional CTA) on performance best-practices: clear value proposition, specificity/numbers, a strong CTA, length discipline, urgency/relevance, and a benefit (not feature) focus. Returns a 0-100 score, per-criterion pass/fail with fixes, and a quick compliance risk flag (delegates depth to compliance_check). With an LLM key it adds two improved variants. Deterministic core. |
| nectarin-intelligence-worker__attribution_model | [NECTARIN Intelligence] Multi-touch attribution simulator. Given conversion PATHS (ordered channel sequences with their conversion counts), it credits conversions to channels under five models — first-touch, last-touch, linear, position-based (U-shaped 40/20/40) and time-decay — and highlights which channels are UNDER- or OVER-valued by naive last-touch vs. multi-touch (the key budget-reallocation insight). Deterministic. |
| nectarin-intelligence-worker__bid_simulator | [NECTARIN Intelligence] Simulate an auction bidding strategy. Using the category's benchmark CPC (derived from CPM/CTR) and conversion rate (from CPC/CPA), it sweeps bid levels and returns a trade-off curve — win-rate, clicks, conversions, spend and resulting CPA at each bid — capped by a daily budget. Recommends the bid that hits a target CPA (if given) or maximizes conversions. Synthetic logistic auction model, clearly labelled. |
| nectarin-intelligence-worker__report_export | [NECTARIN Intelligence] Turn a strategy or analysis into a presentation-ready deck. Pass the `strategy` object (e.g. the structuredContent.data from strategy_orchestrate) and/or your own `sections`; get back ordered slides (title + bullets + speaker notes), a full Markdown deck (--- separated), and a condensed one-pager. Optional LLM polish of the executive summary. Deterministic formatter — compose it after strategy_orchestrate. |
| nectarin-intelligence-worker__localize | [NECTARIN Intelligence] Localize marketing text for RU/CIS markets. Translates and culturally adapts copy into Russian, English, Kazakh or Uzbek while preserving marketing intent and tone. Requires an LLM key; without one it returns the original text with a clear note (never fails). Use it to ship one creative across CIS markets. |
| nectarin-intelligence-worker__creative_testing_matrix | [NECTARIN Intelligence] Multi-variant creative/landing test RESULTS analyzer (the read side of ab_test_planner). From ≥2 arms with OBSERVED visitors + conversions, picks a control (named, else the highest-traffic arm), then for every other arm computes the conversion rate, the absolute & relative lift vs. control, a pooled two-proportion z-test (z, two-tailed p) and significance under a multiple-comparison-corrected α (Bonferroni or Šidák across k−1 comparisons). For non-significant arms it estimates the ADDITIONAL sample per arm needed to detect the observed effect at the target power. Declares WINNER / LOSER / KEEP TESTING / INSUFFICIENT DATA per arm, names the best arm and a roll-out recommendation. Deterministic frequentist statistics on YOUR data — decision support, not a guarantee. |
| nectarin-intelligence-worker__creative_variants | [NECTARIN Intelligence] Generate AND score multiple ready-to-test ad variants for a product × audience × channel. With an LLM key it writes N on-brand, RU-compliant variants (KV-cached); without a key it returns strong deterministic template variants. Every variant is scored by the same heuristic as creative_score (0-100 + grade) and gets a quick compliance flag, then ranked best-first. Pairs with ab_test_planner to test the winners. |
| nectarin-intelligence-worker__anomaly_detector | [NECTARIN Intelligence] Flag anomalies in a metric time series (e.g. daily CPA, CTR, spend, conversions) for always-on monitoring. Uses a robust median/MAD z-score (resistant to outliers), with a std-based fallback for low-variance series. Reports each anomaly's index, value, z-score, direction and severity, whether the LATEST point is anomalous, and the baseline. Deterministic. |
| nectarin-intelligence-worker__cohort_ltv | [NECTARIN Intelligence] Project a cohort's lifetime value from a retention curve. Provide either an explicit retentionCurve (fraction surviving each period, period 0 = 1.0) OR a monthlyChurnPct + periods to synthesize one. Returns per-period survivors/revenue, cumulative LTV per customer and for the whole cohort, optional NPV discounting, and payback period if CAC is given. Complements unit_economics. Deterministic; figures illustrative. |
| nectarin-intelligence-worker__utm_builder | [NECTARIN Intelligence] Build a consistent, validated UTM tracking URL. Normalizes source/medium/campaign/term/content to a chosen casing convention (lower/snake/kebab/preserve), URL-encodes safely, preserves any existing query params, and warns about common mistakes (uppercase, spaces, non-ASCII/Cyrillic, missing required fields). Also returns a campaign naming-convention suggestion. Deterministic; no network call. |
| nectarin-intelligence-worker__pacing_monitor | [NECTARIN Intelligence] Monitor budget pacing against an even spend curve. Given total budget, total/elapsed days and spend-to-date, it computes expected spend, pace ratio, status (under/on-track/over), projected end-of-period spend, remaining budget/days and the recommended daily spend to land exactly on budget. Deterministic; figures illustrative. |
| nectarin-intelligence-worker__response_curve | [NECTARIN Intelligence] Channel saturation / diminishing-returns modeling + budget reallocation. Fits a constant-elasticity response curve (conversions = a·spend^b, 0<b<1) to YOUR current per-channel spend & conversions, then computes the conversion-maximizing split for a target total budget (closed-form: share ∝ a^(1/(1-b))). Returns recommended spend per channel, projected conversions, marginal CPA, blended-CPA improvement and uplift vs. current. Model-based decision support (not real benchmarks). |
| nectarin-intelligence-worker__budget_pacing_forecast | [NECTARIN Intelligence] Trend-aware budget pacing forecast. From the total budget, total/elapsed days and spend-to-date — and optionally the recent daily spend series — it projects end-of-flight spend from the RECENT run-rate (not just a flat average), the over/under-spend variance %, the days to exhaust the budget at the current rate, and the recommended daily rate (and % adjustment) to land exactly on budget. Optional conversions-to-date adds a CPA pace. Complements pacing_monitor (linear) with a trend-based projection. Deterministic; figures illustrative. |
| nectarin-intelligence-worker__utm_taxonomy_qa | [NECTARIN Intelligence] Batch UTM / taxonomy governance auditor. Give it a list of tagged URLs (or raw UTM query strings) and it parses every link, checks each for missing required params (utm_source/medium/campaign by default), uppercase, spaces and non-ASCII/Cyrillic, then aggregates a 0–100 consistency score, near-duplicate value variants per parameter (e.g. 'facebook' vs 'Facebook' vs 'fb'), values outside an optional allow-list for source/medium, and concrete fixes. Complements utm_builder (which builds ONE link) by auditing a whole campaign export for consistency. Deterministic; no network call. |
| nectarin-intelligence-worker__mmm_optimize | [NECTARIN Intelligence] Marketing Mix Model (MMM-lite). From each channel's spend & conversions TIME SERIES, fits adstock/carryover (geometric decay λ, grid-searched by fit R²) and saturation (conversions = a·effectiveSpend^b, log-log least squares, 0<b≤1), then computes the conversion-maximizing STEADY-STATE budget split across channels via exact Lagrange bisection (marginal CPA equalized across funded channels). Returns per-channel adstock decay, saturation elasticity, fit R²/confidence, recommended spend, projected steady-state conversions, marginal CPA, and uplift vs. current. Uses YOUR real series — deterministic, model-based decision support (not real benchmarks). |
| nectarin-intelligence-worker__gtm_calendar | [NECTARIN Intelligence] Build a phased go-to-market roadmap for RU/CIS: splits the horizon into Test → Scale → Optimize phases, assigns goal-driven budget weights and channel emphasis per phase, then produces a WEEK-BY-WEEK budget pacing curve that leans spend into high-demand weeks using the category's monthly seasonality index. Returns per-phase objectives, KPIs and exit criteria, seasonal windows (peak/soft) inside the horizon, and milestones. Deterministic; mock seasonality/benchmarks; not legal advice. Pairs with media_plan / budget_optimizer (what to spend where) — this answers WHEN and in what sequence. |
| nectarin-intelligence-worker__scenario_planner | [NECTARIN Intelligence] What-if budget scenario comparator. Takes CURRENT per-channel spend & conversions plus named scenarios (e.g. conservative/base/aggressive, via a budgetMultiplier and/or absolute per-channel spend overrides) and projects each scenario's conversions, blended CPA, incremental conversions vs. today, and — if revenuePerConversion is given — revenue, profit, ROAS and ROI%. Each channel uses a constant-elasticity diminishing-returns curve conversions=conv₀·(spend/spend₀)^b calibrated to its own current point (b default 0.7, per-channel overridable). Ranks scenarios by objective (max_conversions | min_cpa | max_roi), recommends one with a rationale and an elasticity-sensitivity note. Uses YOUR numbers — deterministic decision support, not benchmarks. Complements mmm_optimize (optimal split) / budget_optimizer (single-budget allocation): this compares YOUR candidate plans head-to-head. |
| nectarin-intelligence-worker__promo_planner | [NECTARIN Intelligence] Promo / discount P&L and break-even calculator. From regular price, variable unit cost and baseline period volume, computes the post-discount unit margin, the BREAK-EVEN volume uplift the promo must clear to avoid losing money, and — if expectedUpliftPct is given — projected units/revenue/profit, incremental profit vs. baseline and ROI on the discount investment. Supports an optional fixed promo cost (media/ops) and a pull-forward/cannibalization penalty on incremental volume. Returns a verdict (profitable / needs more uplift / margin-destroying). Deterministic trade-marketing math on YOUR numbers — decision support, not a guarantee. |
| nectarin-intelligence-worker__price_optimizer | [NECTARIN Intelligence] Profit-maximizing price finder. From ≥2 historical (price, units) observations, fits a constant-elasticity demand curve Q = a·P^(-e) by log-log least squares, estimates the price elasticity of demand, and — when demand is elastic (e>1) — computes the profit-maximizing price P* = cost·e/(e−1) (standard markup rule), with projected units/revenue/profit and the uplift vs. an optional currentPrice. Flags inelastic demand (e≤1, no interior optimum) and low-confidence fits. Deterministic, on YOUR data — decision support, not a guarantee. Complements promo_planner (which evaluates a fixed discount). |
| nectarin-intelligence-worker__marketing_audit | [NECTARIN Intelligence] Senior-level account health audit. Give the current per-channel spend & conversions for a category and NECTARIN scores each channel's CPA against RU/CIS benchmarks (p25/p50/p75), flags concentration risk (one channel hogging budget) and untracked channels (no conversions), computes an overall health score (0-100) + grade A–D, and returns a PRIORITIZED action plan — including a concrete budget reallocation with projected extra conversions and saved spend. Optional targetCpa compares blended CPA to your business goal. Deterministic; benchmarks are mock unless real data is layered into KV. |
| nectarin-intelligence-worker__landing_cro_audit | [NECTARIN Intelligence] Heuristic conversion-rate-optimization (CRO) audit for a landing page. Scores up to seven UX/performance dimensions you provide — page speed (loadTimeSec), bounce rate, mobile parity (mobile vs overall CR), form friction (formFields + stepsToConvert), CTA clarity (hasClearCta + aboveFoldCta), trust & social proof (hasSocialProof + hasTrustSignals), and CR vs an industry benchmark — into a weighted 0-100 CRO score with a letter grade. Returns a prioritized issue list (by weight × gap), concrete fixes, and a projected CR uplift (multiplicative, with diminishing returns) that, given monthlyVisitors + AOV, is translated into incremental conversions & revenue. Heuristic decision support — validate with a real A/B test (see ab_test_planner / creative_testing_matrix), not a guarantee. |
| nectarin-intelligence-worker__board_report | [NECTARIN Intelligence] Executive one-pager (orchestrator). Give a category and current per-channel spend & conversions; board_report internally runs marketing_audit (health score, channel verdicts, concentration/untracked risks, prioritized actions) and scenario_planner (a +15% budget upside scenario), then assembles a board-ready brief: status + grade, headline metrics (spend, conversions, blended CPA, and revenue/profit/ROI if revenuePerConversion is given), best/worst channel, live risks, top recommendations, budget upside and a single next step. Composes the deterministic sub-tools — consistent with marketing_audit / scenario_planner. Decision support on your own numbers, not a guarantee. OPTIONAL: set `useTenantData:true` to fold REAL tenant figures (spend/revenue/conversions/blended CPA/ROAS from the funnel+kpis blob in KV) into the executive summary (mode=kv only; strictly opt-in — absent/mock ⇒ unchanged behavior). |
| nectarin-intelligence-worker__creative_fatigue | [NECTARIN Intelligence] Creative burnout detector. From each creative's daily CTR series (ctr[] in %, or impressions[]+clicks[]), finds peak CTR, decline from peak, the recent least-squares trend, a 0–100 fatigue score + stage (fresh/maturing/fatigued/burnt), and — when CTR is still falling — the estimated days until it crosses the refresh threshold (default 70% of peak). Ranks creatives worst-first and recommends which to refresh now / prepare to refresh / monitor. Deterministic, on YOUR performance series — decision support, not a guarantee. |
| nectarin-intelligence-worker__creative_rotation | [NECTARIN Intelligence] Creative rotation optimizer (portfolio allocation against fatigue). From a set of creatives — each with a performance metric (CTR% or CVR%) and the impressions already served — it applies an exponential fatigue decay (effectiveness halves every `halfLifeImpressions`), then water-fills the next period's impressions to the creatives with the highest fatigue-adjusted value, capped per creative (default 40%) to preserve rotation/variety. Returns the recommended impression share per creative, decay multiplier, status (scale / maintain / retire), the projected aggregate performance uplift vs. an even rotation, and how many fresh creatives to produce. Complements creative_fatigue (single-creative time-series). Deterministic — decision support, not a guarantee. |
| nectarin-intelligence-worker__influencer_planner | [NECTARIN Intelligence] Influencer / KOL roster evaluator & mix optimizer for Маркетинг влияния. For each blogger (followers, price, optional avgViews, ER%, audienceMatch%) computes reach (avgViews, or followers×reachRate), CPM, CPV, CPE, estimated target reach & conversions, eCPA and a value score; FLAGS suspicious engagement (likely inflated/bot or dead audience vs. typical band for the follower tier). When a budget is given, greedily selects the best mix (by eCPA for conversions or CPM for reach) and reports blended reach/conversions/CPA/CPM. Deterministic, on YOUR roster + assumptions — decision support, validate with a test placement. |
| nectarin-intelligence-worker__reach_frequency | [NECTARIN Intelligence] Reach & frequency planner for OLV (online video) and display. From a budget + CPM (or impressions directly) and the target audience universe, computes gross impressions, GRPs, NET reach (people & %), average frequency among reached, the full contact distribution, and EFFECTIVE reach at ≥N exposures using a Poisson exposure model. With an optional frequencyCap it estimates impressions wasted above the cap and the potential reach gain from reallocating them, plus cost-per-reached-person and an under-/over-frequency verdict. Deterministic media math on YOUR plan inputs — a planning estimate, not a guarantee. |
| nectarin-intelligence-worker__channel_overlap | [NECTARIN Intelligence] Omnichannel deduplicated reach estimator. Given a shared audience universe and ≥2 channels' individual reach (reachPct of universe, or reachPeople), computes the combined NET deduplicated reach under the independence (Sainsbury) model, the gross summed reach, the duplication/overlap (people & %), and each channel's incremental UNIQUE reach (leave-one-out) — i.e. how much net reach it adds on top of the others. Flags the most additive and most duplicated channels. Deterministic planning estimate (assumes random duplication) — pair with reach_frequency for single-channel R&F. |
| nectarin-intelligence-worker__media_flowchart | [NECTARIN Intelligence] Media flighting / flowchart planner. Distributes a total budget across N weeks by a flighting pattern (even / front_loaded / back_loaded / burst / pulse), returning the per-week budget, share and cumulative spend, plus a per-channel split each week when channel shares are given. Reports the peak week and on-air weeks. Deterministic scheduling math on YOUR plan — a planning artifact, not a guarantee. |
| nectarin-intelligence-worker__media_quality_score | [NECTARIN Intelligence] Media delivery quality scorer. From a placement's OWN delivered metrics (viewability %, invalid/bot traffic %, video completion %, brand-safe %, in-geo/on-target %), computes a weighted 0–100 quality score and an A–F grade, scores each metric vs. RU market thresholds (MRC-style), flags problems (low viewability, high IVT, weak completion/brand-safety), and gives a verdict + the biggest lever. Complements supplier_quality (which is a benchmark lookup) — this scores YOUR actual delivery. Deterministic, decision support, not a guarantee. |
| nectarin-intelligence-worker__audience_overlap | [NECTARIN Intelligence] Audience overlap / deduplication analyzer from MEASURED pairwise overlaps (e.g. from a DMP, panel or cross-device graph) — unlike channel_overlap, which assumes statistical independence. Given segment sizes (reach % or absolute users) and the measured pairwise overlaps, it computes the deduplicated total reach (inclusion–exclusion), the duplication rate (wasted double-counting), each segment's incremental (leave-one-out) unique contribution and redundancy, a duplication matrix, and which segment is most additive vs. most redundant — to cap frequency or reallocate budget. Exact for 2 segments; a clamped 2nd-order estimate for ≥3 (no triple-intersection data). Deterministic. |
| nectarin-intelligence-worker__frequency_cap_optimizer | [NECTARIN Intelligence] Frequency-cap optimizer for OLV / display. From a fixed impression pool (impressions, or budget + CPM) and the target audience universe, it (A) DIAGNOSES how many impressions land on people already past each candidate cap at the natural average frequency (wasted over-cap impressions, Poisson model), and (B) OPTIMIZES: for each cap it re-solves the per-person delivery so the freed impressions are reallocated, returning the resulting NET (1+) reach, EFFECTIVE reach at ≥N exposures, average frequency and the reach uplift vs. no cap. Recommends the cap that maximises ≥N effective reach. Deterministic media math on YOUR plan inputs — a planning estimate, not a guarantee. |
| nectarin-intelligence-worker__brand_lift | [NECTARIN Intelligence] Brand-lift study calculator for Брендинг. MEASURE mode: from a control vs. exposed survey cell (n + positive answers for a brand metric — ad recall / awareness / consideration / intent) computes both rates, the absolute (pp) and relative lift, a pooled two-proportion z-test (z, two-tailed p-value, significance at α) and a confidence interval for the absolute lift. DESIGN mode: from a base rate + target lift (absolute pp or relative %), α and power, returns the required sample size PER CELL and total. Auto-detects mode from inputs. Deterministic survey statistics on YOUR numbers — decision support, not a guarantee. |
| nectarin-intelligence-worker__sov_tracker | [NECTARIN Intelligence] Share of Voice tracker with ESOV → market-share growth prediction (Binet & Field rule). From the brand's spend + competitors' spends (or a given totalMarketSpend / sovPct) and the brand's current market share (SOM), computes Share of Voice (SOV), Excess Share of Voice (ESOV = SOV − SOM, pp) and the predicted annual market-share growth (~0.5pp per 10pp ESOV by default). Also solves the SOV required to hit a target share growth. Deterministic brand-growth heuristic on YOUR numbers — directional, not a guarantee. |
| nectarin-intelligence-worker__share_of_search | [NECTARIN Intelligence] Share of Search tracker — branded search demand as a LEADING indicator of market share (Les Binet). From the brand's branded-search volume + competitors' volumes (or a totalCategoryVolume, or sosPct directly), computes Share of Search (SoS %), the brand's rank, and — given the current market share (SOM) — the SoS↔share GAP (SoS above share ⇒ likely to gain; below ⇒ at risk). Optionally takes a previous SoS to report the trend, and projects next-period market share as the share partially converging toward SoS. Deterministic demand-sensing heuristic on YOUR numbers — directional, not a guarantee, and distinct from sov_tracker (which tracks share of media SPEND). |
| nectarin-intelligence-worker__production_estimator | [NECTARIN Intelligence] Creative production budget & timeline estimator for Производство. From a list of deliverables (asset type × quantity × complexity) and a quality tier (economy/standard/premium), applies an illustrative RU rate card to give a per-deliverable cost & effort breakdown, a subtotal, contingency and optional rush surcharge, a total cost RANGE (±20%), and a critical-path timeline estimate (production is partly parallel). Asset types: video, video_cutdown, static, key_visual, animated_banner, social_post, photo, landing, audio. Heuristic & deterministic — a planning ballpark, confirm with real vendor quotes. |
| nectarin-intelligence-worker__geo_holdout | [NECTARIN Intelligence] Geo-based incrementality (matched-market holdout) test tool. DESIGN mode: from the expected baseline conversions in the test geos over the window and a target lift, returns the minimum detectable lift (MDE), the baseline volume required to detect the target lift, and a recommended test duration when a weekly baseline is given. MEASURE mode: from observed test-geo conversions vs. a counterfactual (e.g. scaled control geos), computes incremental conversions, lift %, a count-based (Poisson) z-test, two-tailed p-value, significance and the incremental CPA when test spend is given. Auto-detects mode. Deterministic count-data statistics on YOUR numbers — decision support, not a guarantee. |
| nectarin-intelligence-worker__incrementality_meta | [NECTARIN Intelligence] Meta-analysis of multiple incrementality / A-B / geo-holdout tests. Each test contributes a lift (%) with a standard error (or a 95% CI, from which SE is derived). It computes the inverse-variance fixed-effect pooled lift (z, p-value, CI), the heterogeneity statistics Q and I², and the DerSimonian–Laird random-effects pooled lift (which widens the CI when results disagree). Returns per-test weights, both pooled estimates, a heterogeneity verdict and an overall significance call — to combine many small reads into one defensible number. Deterministic statistics on YOUR test results. |
| nectarin-intelligence-worker__competitive_response | [NECTARIN Intelligence] Competitive war-game simulator. Given your spend, the current competitor spend and a competitor move (spend escalation %, a new entrant, or a pullback), it models the impact on your Share of Voice (SOV), auction CPM inflation and effective impressions at a fixed budget — then sizes the defensive budget needed to hold your SOV (or a target SOV) and recommends a response posture (hold / partial match / defend or pivot). Deterministic auction-share dynamics on YOUR numbers — directional decision support, not a guarantee. |
| nectarin-intelligence-worker__search_planner | [NECTARIN Intelligence] Paid-search / SEM keyword-portfolio planner for Yandex Direct & контекст. From keywords (monthly search volume + CPC, optional CTR%/CVR%/intent) and an optional monthly budget, estimates per-keyword clicks, conversions, CPA and the max addressable spend; ranks keywords by efficiency (expected conversions per ₽); greedily allocates the budget to the lowest-CPA keywords first; and returns portfolio totals — clicks, conversions, blended CPA, total spend and demand coverage. Defaults CTR=4%, CVR=2% when missing (flagged). Deterministic media math on YOUR keyword inputs — planning estimate, not a guarantee. |
| nectarin-intelligence-worker__retail_media_planner | [NECTARIN Intelligence] Marketplace / retail-media planner for Ozon, Wildberries, Yandex Market & Avito. From placements (search/catalog/banner) with a cost model (CPC, or CPM+CTR), click→order CVR, an average order value (AOV), the marketplace commission (take-rate %) and an optional budget, it computes per-placement effective CPC, orders, revenue, ДРР (доля рекламных расходов = ad spend / revenue) and ROAS, ranks placements by profit per ₽, greedily allocates the budget to the most profitable placements first (respecting click/impression caps), and returns blended portfolio economics (revenue, ДРР, ROAS, net profit) plus a target-ДРР check. Deterministic retail-media math on YOUR inputs — planning estimate, not a guarantee. |
| nectarin-intelligence-worker__churn_predictor | [NECTARIN Intelligence] Churn & retention economics tool for CRM / lifecycle. Resolves a MONTHLY churn rate from one of: monthlyChurnRatePct, OR a cohort (customersStart + customersRetained over periodMonths), OR monthlyRetentionPct. Then computes annualised churn, average customer lifetime (1/churn), a survival curve to the horizon, customers & revenue retained vs. lost, and LTV (ARPU/churn, optionally discounted). Given a retention initiative (reduceChurnByPp + programCost) it sizes the LTV uplift per customer, the total uplift and the ROI of retention. Deterministic retention math on YOUR numbers — decision support, not a guarantee. |
| nectarin-intelligence-worker__rfm_segmenter | [NECTARIN Intelligence] RFM (Recency–Frequency–Monetary) customer segmentation. From a list of customers with recencyDays + frequency + monetary, it scores each on 1–5 quintiles (recency inverted — more recent = higher), combines R with the F/M average into the classic named segments (Champions, Loyal, Potential Loyalist, At Risk, Can't Lose Them, Hibernating, Lost, …), then sizes every segment (customers, share %, total & average monetary) and attaches a concrete CRM action. Surfaces the revenue concentrated in Champions and the revenue at risk (At Risk + Can't Lose Them). Deterministic segmentation on YOUR data — decision support, not a guarantee. |
| nectarin-intelligence-worker__email_campaign_planner | [NECTARIN Intelligence] Email / CRM newsletter economics & cadence planner. From a list size, deliverability, open rate, click rate (clicks/delivered, or derive from clickToOpen), conversion rate (orders/clicks) and AOV, it computes per-send delivered → opens → clicks → orders → revenue and the key revenue-per-email (RPE). With sendsPerMonth it projects monthly & annual revenue, orders and list attrition from unsubscribes, plus a list half-life and a fatigue warning when cadence × unsubscribe is high. With costPerEmail and/or platformMonthlyCost it returns profit and ROI. Deterministic email math on YOUR numbers — a planning estimate, not a guarantee. |
| nectarin-intelligence-worker__affiliate_program_planner | [NECTARIN Intelligence] CPA / affiliate / partner-program economics planner for RU networks (Admitad, Cityads, …) and direct partners. From AOV, gross margin %, a commission model (percent of AOV via commissionPct, or fixed CPA via cpaPayout), an optional networkFeePct and validationRatePct (approved orders), plus per-partner clicksPerMonth & conversionRatePct, it computes per-partner approved orders, revenue, payout, EPC (partner earnings per click), effective CPA, ROAS and net profit to the advertiser, ranks partners best-first, blends the whole program, and derives the SUSTAINABLE commission ceiling (payout where profit per order = 0 ⇒ margin/(1+fee)). Flags loss-making partners and checks an optional target CPA. Deterministic affiliate math on YOUR numbers — a planning estimate, not a guarantee. |
| nectarin-intelligence-worker__seo_opportunity | [NECTARIN Intelligence] SEO organic-growth opportunity model for an SEO specialist. From a list of keywords (monthlySearchVolume + currentPosition + targetPosition) and a conversion rate + value per conversion (or AOV), it applies a position→CTR curve to estimate current vs. target organic traffic, the incremental clicks, conversions and revenue per keyword, ranks the biggest opportunities, and flags 'quick wins' (page-2 keywords, positions 11–20, that are cheap to push onto page 1). Returns portfolio totals and a verdict. Deterministic SEO math on YOUR keywords — a planning estimate, not a ranking guarantee. |
| nectarin-intelligence-worker__social_media_planner | [NECTARIN Intelligence] Organic social-media / SMM planner for an SMM or community manager. From one or more platforms (VK, Telegram, Дзен, YouTube, …) with followers, postsPerWeek, organic reachRatePct (% of followers reached per post) and engagementRatePct (of reached), it projects monthly posts, reach, impressions, engagements, follower growth (from an optional growthRatePct) and — with conversionRatePct + aov — conversions & revenue. Aggregates the portfolio, recommends a cadence, and flags low organic reach. Deterministic SMM math on YOUR numbers — a planning estimate, not a guarantee. |
| nectarin-intelligence-worker__pr_value_estimator | [NECTARIN Intelligence] PR / earned-media value & share-of-voice estimator for a PR or communications manager. From a list of placements (outlet, audienceReach, optional tier and sentiment) it computes total potential reach (with an overlap discount), a tier- & sentiment-weighted QUALITY-ADJUSTED reach, and an advertising-equivalent reach value using a CPM benchmark (clearly labelled — AVE is a context metric, not an endorsed KPI). With competitorReach it computes earned share of voice. Returns a per-placement table, a PR quality score and a verdict. Deterministic on YOUR data — context, not a guarantee. |
| nectarin-intelligence-worker__event_roi_planner | [NECTARIN Intelligence] Event / webinar / field-marketing ROI planner. Projects the full funnel from an audience (invites or reach) through registrations → attendees → leads → opportunities → won deals → revenue using the rates you provide, then computes cost per registration / attendee / lead, pipeline value, ROI and a breakeven (deals or revenue needed to cover the cost). Works for webinars, conferences, expos and field events. Deterministic on YOUR numbers — a planning estimate, not a guarantee. |
| nectarin-intelligence-worker__aso_planner | [NECTARIN Intelligence] App Store Optimization (ASO) & mobile-growth planner. From store impressions, the tap-through rate (impression→product page view) and the install conversion rate (page view→install), it projects page views and installs; with d1/d30 retention + ARPDAU or LTV it estimates revenue, and with a paid CPI it sizes paid user acquisition economics and breakeven. Includes an ASO uplift scenario (improve the install conversion rate by N pp → extra installs & value). Deterministic mobile math on YOUR numbers — a planning estimate, not a guarantee. |
| nectarin-intelligence-worker__content_plan_roi | [NECTARIN Intelligence] Content-marketing ROI model that treats content as a COMPOUNDING asset. Given pieces produced per month, cost per piece, the steady-state monthly visits each piece earns (after a ramp), an optional content lifespan, plus a conversion rate and value per conversion, it simulates the library month-by-month over a horizon: cumulative published pieces, total monthly & cumulative organic visits, conversions, value, content spend, net ROI and the payback month. Shows why content pays back later but compounds. Deterministic on YOUR numbers — a planning estimate, not a guarantee. |
| nectarin-intelligence-worker__role_playbook | [NECTARIN Intelligence] Role router & adoption engine: maps a marketer's profession to their tailored NECTARIN toolkit so EVERY specialist (not just the CMO) has a first-class reason to use the connector. Pass a role in RU or EN (free-text, alias-matched) — e.g. 'SEO', 'таргетолог', 'CRM', 'медиапленер', 'PR', 'бренд-менеджер', 'аналитик', 'маркетплейсы' — to get the role's primary tools (with a role-specific reason for each), the supporting toolkit, an ordered end-to-end workflow, the KPIs that role owns, and example questions. Call WITHOUT a role to list every supported profession (for leadership / onboarding). Curated, deterministic mapping — all access flows through one Unyly install. |
| nectarin-intelligence-worker__connect_via_unyly | [NECTARIN Intelligence] FRONT DOOR — how to connect / install / upgrade NECTARIN Intelligence, always through Unyly. Returns the tracked Unyly install link (unyly.org listing, with UTM attribution by source/role/plan), the manual MCP endpoint as a fallback, role-aware onboarding (pairs with role_playbook) and the access tiers (free / pro / team / agency) with what each unlocks. Use this whenever a user asks how to get access, add the connector, onboard their team, or upgrade — so consumption, requests and installs flow through Unyly (the metering & governance point; Unyly Connect fronts OAuth 2.1). Returns links & guidance only — no PII, no network call. |
| nectarin-intelligence-worker__marketing_skill | [NECTARIN Intelligence] SKILLS / playbooks engine — composable, end-to-end marketing workflows that chain several NECTARIN tools into one repeatable job. With NO arguments it lists the catalogue of skills (launch, cut CAC, retention, creative refresh, budget reallocation, SEO/content, social/influencer, board readout, marketplace scaling, measurement setup). Given a `skill` name/alias OR a free-text `goal` (RU/EN), it returns the ordered recipe: which tools to run, in what order, why, the inputs needed and the KPIs to watch. Deterministic planning only — it returns the workflow, it does NOT call the tools itself. Extensible: this is the layer to add new repeatable playbooks. |
| nectarin-intelligence-worker__cohort_retention_curve | [NECTARIN Intelligence] Cohort retention modeling for a product / growth / CRM marketer. From YOUR cohort retention points (day + retention %, e.g. D1=40, D7=22, D30=12) it fits a power-law curve r(t)=a·t^(−b) via log-log least squares, projects retention at D1/D7/D30/D90/D365, reports fit quality (R²), and — given ARPU per active user per day — estimates LTV over a horizon (default 365 days) as ARPU×Σr(t). Use it to forecast long-run retention and LTV from a short observed window. Deterministic curve-fit on your data — a planning estimate, not a guarantee. |
| nectarin-intelligence-worker__viral_loop | [NECTARIN Intelligence] Referral / virality model for a growth marketer. From invites per user (i) and invite→signup conversion (c%), it computes the viral k-factor (k=i·c), classifies the loop (viral if k≥1), and for k<1 the amplification multiplier 1/(1−k). Given a paid/seed cohort it projects total users (seed×amplification) and the organic uplift. Optional referral-incentive economics: profit per referred user (LTV−incentive) and the break-even incentive ceiling. Deterministic growth math on your inputs — a planning estimate, not a guarantee. |
| nectarin-intelligence-worker__mcp_federation | [NECTARIN Intelligence] FEDERATION / marketplace router — NECTARIN is the marketing hub; the best specialist MCP servers are added around it, always through Unyly. With NO arguments it lists the catalogue of complementary external MCPs (live keyword data, web analytics, ad-platform pulls, creative generation, social listening, CRM data, marketplace data, localization) with what each adds and which native NECTARIN tools it pairs with. Given a `capability`/`goal`/`role` it recommends the right servers; given a `server` key it returns details. Every entry includes a tracked Unyly connect link so installs & consumption flow through unyly.org (the metering/billing point). Discovery + routing + links only — no PII, no network call; runtime proxying is brokered by the Unyly gateway. |
| nectarin-intelligence-worker__federation_invoke | [NECTARIN Intelligence] FEDERATION runtime — call a tool on a federated external MCP server THROUGH NECTARIN (the hub). Fail-closed: reaches out when `UNYLY_GATEWAY_TOKEN` is set (calls `gateway.unyly.org/mcp/<slug>` with one token for all servers) or when per-server `FED_<KEY>_URL` is set (override). You pick a `server` from the mcp_federation catalogue and the external `tool` name + `arguments`; NECTARIN proxies a single JSON-RPC tools/call and returns the result. If neither gateway token nor per-server URL is configured, it returns a tracked Unyly connect link and makes NO network call. No arbitrary URLs (only known registry keys) — traffic flows through Unyly-brokered endpoints. |
| nectarin-intelligence-worker__marketing_maturity_assessment | [NECTARIN Intelligence] Marketing maturity scorecard for a CMO / head of marketing / transformation lead. Rate 7 capability dimensions 0–5 (strategy, data, measurement, channels, martech, team, creative) and it computes a weighted 0–100 maturity index, the maturity LEVEL (1 Nascent → 5 Leading), per-dimension strengths vs. gaps (with the weighted shortfall to 'leading'), and a prioritized 90-day roadmap targeting the highest-leverage gaps first. Provide any subset of dimensions; unrated ones are reported as not assessed. Deterministic weighting on your self-assessment — a planning compass, not an audit of a live account (use marketing_audit / account_audit for that). |
| nectarin-intelligence-worker__martech_stack_roi | [NECTARIN Intelligence] MarTech stack ROI & rationalization for a marketing-ops / RevOps lead. From your tools (name, annualCost ₽, utilizationPct 0–100, category, optional satisfaction 1–5) it computes total annual spend, wasted spend (cost × idle share), category redundancy (multiple tools in one category ⇒ keep the best-utilized, flag the rest), low-utilization cut candidates (<30%), projected consolidation savings, and a utilization-weighted ROI of the stack. Returns a ranked rationalization plan. Deterministic accounting on your inputs. |
| nectarin-intelligence-worker__pricing_psm | [NECTARIN Intelligence] Van Westendorp Price Sensitivity Meter (PSM) for product / pricing research. From survey respondents — each giving four prices: tooCheap (so cheap quality is doubted), cheap (a bargain), expensive (starting to be pricey but worth considering), tooExpensive (would not buy) — it builds the four cumulative curves and locates the OPP (Optimal Price Point), IPP (Indifference Price Point), and the acceptable price band PMC→PME (points of marginal cheapness/expensiveness). Respondents with non-monotonic prices are dropped and reported. Deterministic intersection of empirical curves on your data. |
| nectarin-intelligence-worker__abm_account_scoring | [NECTARIN Intelligence] ABM / B2B account prioritization for a demand-gen or account-based marketer. For each target account give fit (ICP match 0–100), intent (buying signals 0–100) and engagement (your touch/response 0–100), optional dealSize ₽. It computes a weighted 0–100 priority score (default weights fit .40 / intent .35 / engagement .25, overridable), assigns a tier (1:1 / 1:few / 1:many / nurture) with a recommended play, and — when dealSize is given — an expected-value ranking (score × dealSize). Deterministic weighting on your inputs. |
| nectarin-intelligence-worker__nps_analysis | [NECTARIN Intelligence] Net Promoter Score (NPS) analysis for a CX / customer-marketing / loyalty owner. Provide either raw 0–10 `scores` or aggregate `counts` {promoters, passives, detractors}. It returns the segment split (promoters 9–10, passives 7–8, detractors 0–6), the NPS (−100..+100), a 95% confidence interval (NPS standard error), and a benchmark interpretation band. Deterministic survey math. |
| nectarin-intelligence-worker__b2b_pipeline_velocity | [NECTARIN Intelligence] B2B sales/pipeline velocity for a revenue / demand-gen marketer. Velocity = (qualified opportunities × win-rate × average deal size) ÷ sales-cycle length (days) = revenue generated per day. Returns daily/monthly/annual velocity and a +10% lever sensitivity (opportunities, win-rate, deal size, and a −10% on cycle length) to reveal the highest-leverage improvement. Deterministic formula on your funnel numbers. |
| nectarin-intelligence-worker__win_loss_analysis | [NECTARIN Intelligence] B2B win/loss analysis for a revenue / product-marketing team. From closed deals (outcome won|lost, optional reason, segment, value ₽) it computes the overall win rate (by count and by value), win rate by segment, the top loss reasons and top win reasons (count + value impact), and prioritized recommendations targeting the biggest loss drivers. Deterministic aggregation on your CRM export. |
| nectarin-intelligence-worker__kpi_alert_engine | [NECTARIN Intelligence] Cross-KPI rule-based ALERT ENGINE — the autonomy layer that turns a dashboard into a prioritized to-do list. For each metric give value and a target (or benchmark), plus direction ('higher_better' or 'lower_better'; inferred from the name when omitted — CPA/CAC/churn ⇒ lower-better). It grades each KPI ok/watch/warning/critical by adverse deviation (default warn 10%, crit 25%), and on every breach maps it to a recommended ACTION and the NECTARIN tool to run next (e.g. CPA↑ ⇒ budget_optimizer, CTR↓ ⇒ creative_testing_matrix, churn↑ ⇒ churn_predictor). Returns alerts sorted by severity. Deterministic anomaly→action routing. OPTIONAL: set `useTenantData:true` to auto-populate current KPI values (and targets) from the tenant metrics blob in KV (mode=kv only; strictly opt-in — absent/mock ⇒ unchanged behavior, provide `metrics` manually). |
| nectarin-intelligence-worker__marketing_budget_allocator | [NECTARIN Intelligence] CMO annual marketing-budget allocator ACROSS FUNCTIONS (not media channels — for channel splits use budget_optimizer). Splits a total budget across brand, demand/performance, retention/CRM, content/SEO, martech and team/ops using benchmark shares tilted by your primary goal (awareness | growth | performance | efficiency | retention), clamped to sensible guardrails per function. Returns ₽ + % per function, the tilt applied vs. the benchmark, and guardrail notes. Deterministic; a starting framework to negotiate, not a mandate. |
| nectarin-intelligence-worker__autonomous_plan | [NECTARIN Intelligence] AUTONOMY: turns a set of KPI breaches OR a goal into an ORDERED, deterministic remediation/action PLAN that chains existing NECTARIN tools. Give `breaches` (each: name, optional severity critical|warning|watch, optional deviationPct) and/or a free-text `goal`. It assembles a coherent multi-step plan: a DIAGNOSE step, then one ACTION step per breach (severity-first; each names the issue, the recommended NECTARIN tool, the owner/role, the expected-impact direction and what it depends on), then a MEASURE/control step — and for a goal it points to the matching `marketing_skill` recipe. Complements `kpi_alert_engine` (single-alert routing) by sequencing alert→action→measure. Deterministic; returns the plan/recipe, it does NOT execute other tools. |
| nectarin-intelligence-worker__benchmark_kpi_check | [NECTARIN Intelligence] AUTONOMY (Phase D): compare a KPI value against RU/CIS benchmark bands (p25/p50/p75) instead of a manual target. Give category, kpi (CPM|CTR|CPA|VTR or alias), current value and optional platform — it pulls bands from the same data layer as ru_benchmarks, grades ok/watch/warning/critical vs. the spread, and routes breaches to a recommended action + NECTARIN tool (ACTION_MAP). Complements kpi_alert_engine (manual targets) for benchmark-native alerting. |
| nectarin-intelligence-worker__alert_to_skill | [NECTARIN Intelligence] AUTONOMY (Phase D): turns KPI breaches (output of kpi_alert_engine / benchmark_kpi_check) into an ORDERED workflow — alert summary → matched marketing_skill recipe → tool chain → first tool to call. Give `alerts` [{name, severity?, suggestedTool?, deviationPct?}] and/or raw `metrics`. Deterministic skill matching via the skills catalogue (matchSkill); does NOT execute tools. Complements autonomous_plan (generic remediation) with playbook-native routing. Pro+. |
| nectarin-intelligence-worker__autonomous_execute | [NECTARIN Intelligence] AUTONOMY (Phase D) — the GUARDED alert→action→measure runner: moves from planning to (safe) execution. Give an ordered recipe as `steps` (each: intent/action, and EITHER a local NECTARIN `tool` for a read-only step OR a federated `server`+`toolName` for an external WRITE step), and/or `breaches`/`goal` to derive read-only steps (same routing as autonomous_plan). SAFETY MODEL: `mode` defaults to 'dry_run' — returns exactly what WOULD be done (ordered steps, targets, arguments) with ZERO side effects, deterministic like a plan. `mode:'execute'` additionally requires `confirm:true` + a non-empty `approvalToken`; otherwise it degrades to the dry-run plan. WRITE/external steps run ONLY through federation_invoke (fail-closed): with no UNYLY_GATEWAY_TOKEN / per-server override no network call is made and the step is blocked with a 'connect via Unyly' note. Read-only NECTARIN steps are described, never auto-run (no recursive fan-out). Team+ tier. Returns a full audit trail (per step: intent, t |
| nectarin-intelligence-worker__marketing_okr_planner | [NECTARIN Intelligence] Marketing OKR planner for a CMO / head of marketing / team lead. Give a qualitative `objective` and 1–7 `keyResults` (each: metric, baseline, target, optional unit/direction). It returns a measurable OKR: per key result the baseline → target, absolute delta, % change, an AMBITION band (conservative / realistic / ambitious / stretch by magnitude), a LEADING-vs-LAGGING classification (outcome metrics like revenue/CAC/NPS = lagging; activity metrics like traffic/leads/CTR = leading) and the NECTARIN tool to drive it. Warns when the set is unbalanced (all-lagging or all-leading). Deterministic OKR math on your numbers. |
| nectarin-intelligence-worker__content_calendar_planner | [NECTARIN Intelligence] Content-team CAPACITY / throughput planner for a content lead / SMM manager / editor. From `people`, productive `hoursPerWeek` per person, a planning horizon in `weeks`, and a content mix (`contentTypes`: each type with effort-hours per piece, an optional desired share weightPct and an optional `planned` target count) it computes total capacity hours, the achievable pieces per type, weekly throughput, utilization vs. the requested plan and the bottleneck (the type whose demand most exceeds capacity). Deterministic capacity math — plan realistic editorial calendars, not wishful ones. |
| nectarin-intelligence-worker__demand_forecast | [NECTARIN Intelligence] Deterministic demand / time-series FORECAST for any marketing series (sales, leads, traffic, revenue, installs). Give an ordered `series` of ≥3 equally-spaced observations and it projects the next `periods` using Holt's linear-trend double exponential smoothing — returning a point forecast, a residual-based confidence band (lower/upper from in-sample one-step RMSE, widening with the horizon), the fitted level & per-period trend, MAPE and fit quality. Optionally pass `seasonLength` (e.g. 12 monthly, 7 daily, 4 quarterly) with ≥2 full cycles of data to add MULTIPLICATIVE seasonality via classical decomposition (deseasonalize → Holt → reseasonalize). Distinct from seasonality_forecast (category index) and anomaly_detector (outlier flagging). Deterministic math on your numbers — a projection, not a guarantee. |
| nectarin-intelligence-worker__customer_journey_map | [NECTARIN Intelligence] Customer-journey / lifecycle map for a CMO, lifecycle or CRM marketer. Maps the funnel (awareness → consideration → purchase → retention → advocacy) to the channels, content and primary KPI for each stage, computes stage-to-stage CONVERSION when you pass volumes (`count`), and flags COVERAGE GAPS — any stage with no channels or no content planned — plus the biggest drop-off. Call with no args to get the best-practice template, or pass `stages` (each: stage name + optional channels[]/content[]/kpi/count) to map YOUR setup and audit it. Deterministic gap analysis on your inputs. |
| nectarin-intelligence-worker__competitive_positioning_map | [NECTARIN Intelligence] 2-axis competitive POSITIONING / perceptual map for a brand strategist or CMO. Give ≥2 `competitors` each with an `x` and `y` score (e.g. x = price, y = value/quality; any 0–100 or absolute scales) and it places everyone on the plane, splits at the MEAN of each axis into four quadrants, assigns each player a quadrant, computes a value-for-money index (y/x), finds the empty quadrants (WHITE-SPACE opportunities), and — if one entry is flagged `isYou` — reports your quadrant and your nearest rival by normalized distance. Label the axes via `xAxis`/`yAxis`. Deterministic positioning math on your inputs. |
| nectarin-intelligence-worker__marketing_roi_waterfall | [NECTARIN Intelligence] Period-over-period REVENUE / ROAS waterfall (bridge) for a CMO or marketing analyst. Pass two periods (`before` and `after`), each with spend + conversions + revenue, and it decomposes the change in revenue into THREE exact, reconciling drivers via sequential factor substitution: Revenue = Spend × Efficiency (conversions per ₽) × AOV (revenue per conversion). Returns a waterfall (start → +spend effect → +efficiency effect → +AOV effect → end) with each driver's signed ₽ contribution and share of the total change, plus the ROAS before/after and its delta. Use it to explain WHY revenue or ROI moved (more budget? cheaper conversions? bigger basket?). Distinct from unit_economics (single-period) and scenario_planner (forward what-ifs). Deterministic exact decomposition on your numbers. |
| nectarin-intelligence-worker__conjoint_analysis | [NECTARIN Intelligence] Choice-based / part-worth CONJOINT analysis (lite) for a product or pricing marketer. Give `attributes` (each with `levels`, each level carrying a part-worth `utility` from a prior conjoint study or expert estimate) and it computes: attribute IMPORTANCE (each attribute's utility range as a share of the total), the OPTIMAL feature bundle (max-utility level per attribute), and — if you pass `profiles` (named bundles, each choosing one level per attribute) — the SHARE OF PREFERENCE across them via a logit (Bradley-Terry-Luce) model. If one attribute is numeric price (`priceAttribute` with numeric level labels), it also estimates the marginal utility of money and the WILLINGNESS-TO-PAY (₽) for each attribute. Distinct from pricing_psm (Van Westendorp, price only) — this trades off features × price. Deterministic utility math on your inputs. |
| nectarin-intelligence-worker__tam_sam_som | [NECTARIN Intelligence] Market SIZING funnel (TAM → SAM → SOM) for a strategist, founder or new-market lead. Two methods: TOP-DOWN — pass `tam` (total market, ₽ or customers) plus `samSharePct` (serviceable share) and `somSharePct` (obtainable share of SAM); or BOTTOM-UP — pass `population` (total potential customers), `penetrationPct` (→ serviceable customers = SAM), `obtainableSharePct` (your realistic share → SOM) and `arpu` (annual revenue per customer) to build sizes in customers AND revenue. Returns TAM/SAM/SOM, the SOM-as-%-of-TAM reality check, and an optional multi-year SOM projection at `cagrPct` over `years`. Light, deterministic market-sizing math — assumptions in, sizes out. |
| nectarin-intelligence-worker__brand_health_index | [NECTARIN Intelligence] Composite BRAND HEALTH index (0–100) for a brand manager / CMO. From funnel-stage inputs — aided awareness, consideration and preference (each 0–100 %) and NPS (−100..+100) — it computes a weighted health score, optional funnel-efficiency sub-score (impression→awareness→consideration→purchase conversion ratios), a benchmark band (Weak / Average / Strong / Leader) and the weakest lever. Distinct from brand_lift (survey lift test) and share_of_search (demand proxy). Deterministic on YOUR numbers. |
| nectarin-intelligence-worker__gtm_launch_readiness | [NECTARIN Intelligence] GTM LAUNCH readiness scorecard for a product marketer / launch lead. Rate five launch pillars 0–5 (product/offer fit, messaging, channels, ops/execution, legal/compliance) and it computes a weighted readiness % (0–100), a go/no-go/conditional verdict, prioritized gaps (pillars <3) and pillar-specific actions. Optional `category` adds regulated-industry compliance notes (pharma, finance). Distinct from marketing_maturity_assessment (company-wide maturity) and gtm_calendar (timing). Deterministic self-assessment. |
| nectarin-intelligence-worker__tenant_metrics_snapshot | [NECTARIN Intelligence] READ-ONLY tenant metrics snapshot — foundation for future data connectors. Given a tenantId, returns the latest metrics JSON blob. With NECTARIN_TENANT_DATA_MODE=kv and NECTARIN_KV bound, reads key `tenant:<id>:metrics`; otherwise returns a demo/mock blob with a clear note. Does NOT write data. Use with kpi_alert_engine / benchmark_kpi_check for autonomous monitoring once real feeds land. |
| nectarin-intelligence-worker__design_system_spec | [NECTARIN Intelligence] DESIGN SYSTEM generator (flagship, deterministic — NOT image generation). From brand inputs (brand, optional industry, mood minimal|bold|premium|playful|corporate, optional primaryColor hex OR baseHue 0–360, optional baseFontSize px) it derives a full design-system SPEC: an HSL-wheel colour palette (primary + analogous secondary + complementary accent + light/dark neutrals, all as hex), a modular type scale (base × ratio → h1..caption), a 4/8pt spacing scale, border-radius + elevation tokens, and a WCAG contrast check for text-on-primary (relative luminance + contrast ratio, AA/AAA pass/fail with the recommended text colour). Same input → same tokens. Rendering of real assets is delegated to a federated creative-generation MCP via Unyly; NECTARIN produces the spec. |
| nectarin-intelligence-worker__banner_layout_spec | [NECTARIN Intelligence] BANNER LAYOUT spec (deterministic, RU platforms — NOT image generation). From a brief (message, optional cta, platform vk|yandex|telegram|ok|web, optional formats filter) it returns per-format banner SPECS: canonical ad sizes for the platform, a layout skeleton per size (logo/headline/visual/CTA zones as % positions with safe margins, orientation-aware), a recommended max headline/CTA text length and a font-size floor for readability at each size. Same input → same specs. Hand the spec to a designer or a federated creative-generation MCP (via Unyly) to render the actual banner. |
| nectarin-intelligence-worker__ad_spec_checker | [NECTARIN Intelligence] AD SPEC / COMPLIANCE linter (deterministic RU ad checks — NOT image generation). Give a creative's attributes (platform vk|yandex|telegram|ok|web, optional textLength, hasOrdLabel bool for ORD/ЕРИР 'erid' marking, textOverImagePct, imageRatio e.g. '16:9', ctaPresent bool, fileWeightKB, regulatedCategory bool) and it lints them against platform + RU ad-law rules: ORD/ЕРИР labelling, text-over-image %, format ratio, file weight, CTA presence, and a regulated-category legal gate. Returns pass/warn/fail per rule, an overall verdict and prioritized fixes. Deterministic; not legal advice — regulated categories still require legal sign-off. |
| nectarin-intelligence-worker__programmatic_trading_plan | [NECTARIN Intelligence] PROGRAMMATIC / DSP trading plan (deterministic — programmatic mechanics, NOT net-reach planning). For a DSP/programmatic trader: from budget (RUB), an optional targetCpm and/or targetCpa, a deal-type mix across PMP / open exchange / PG (programmatic guaranteed), a target viewability % and an optional frequency cap and flight length, it computes a per-deal-type bid strategy, budget pacing (daily + per deal type), expected impressions and de-duplicated reach, and the viewability-adjusted EFFECTIVE CPM (spend / viewable impressions × 1000) versus the viewability goal. Same input → same plan. Distinct from reach_frequency (net reach/frequency) and media_flowchart (flighting): this models programmatic buying mechanics. Planning heuristics, not a bid guarantee. |
| nectarin-intelligence-worker__marketplace_seo_plan | [NECTARIN Intelligence] MARKETPLACE listing-SEO plan (deterministic — ORGANIC ranking, NOT ads). For a Wildberries / Ozon / Yandex Market specialist: from category, listing attributes (title, imagesCount, hasVideo, richContent, attributesFilledPct, reviewsCount, rating, priceCompetitivenessPct), a set of search terms (with optional monthlyVolume & currentPosition) and competition level, it returns a 0–100 listing-optimization score, a keyword→slot mapping (which term belongs in title vs attributes vs description by volume), a ranking-factor checklist (pass/warn/fail + fix), a projected organic position per term and the expected weekly-visibility uplift (impressions from volume × position-CTR). Same input → same plan. Distinct from retail_media_planner (paid ДРР/ROAS) and seo_opportunity (web search). Planning heuristics, not the marketplace's real ranking algorithm. |
| nectarin-intelligence-worker__lifecycle_flow_planner | [NECTARIN Intelligence] LIFECYCLE / CRM automation FLOW planner (deterministic — cross-channel triggered blueprint, NOT a single blast). For a lifecycle / marketing-automation specialist: given a flowType (welcome | onboarding | winback | abandoned_cart | reactivation), the monthly audience entering the flow, an optional AOV and an optional channel allow-list, it designs the triggered flow — ordered stages with delay (hours), channel, purpose and per-step conversion — then computes the funnel (per-step + cumulative conversions), the overall flow conversion rate and a monthly revenue projection (conversions × AOV). Same input → same blueprint. Distinct from email_campaign_planner (one campaign's economics) and rfm_segmenter (base cuts): this is the FLOW/automation design. Planning heuristics, not guaranteed rates. |
| nectarin-intelligence-worker__plg_activation_model | [NECTARIN Intelligence] PLG activation model (deterministic — product-led SaaS activation funnel). For a PLG/growth SaaS marketer: from monthly signups, an aha-moment metric label, baseline signup→activated and activated→retained rates, and a set of activation levers (onboarding_checklist, time_to_value, empty_states, activation_email, in_app_guidance, social_proof), it models the activation funnel (signup → activated → retained), applies each lever's additive activation-rate uplift (capped), and projects activated & retained users plus optional MRR (from ARPU). Same input → same model. Distinct from viral_loop (k-factor), funnel_model (generic funnel) and cohort_retention_curve (retention shape): this is ACTIVATION-rate specifically. Planning heuristics, not guaranteed rates. |
| ara__generate_image | [ARA] Generate image(s) from a text prompt (and optional reference images). Returns result URLs. This tool waits for the result and usually returns the final URL directly. If instead it returns a generation_id with "NOT FINISHED", the job is still running: call generation_status repeatedly until it returns "Completed", then show the user the result. Never tell the user the generation merely started — finish it first. |
| ara__generate_video | [ARA] Generate a video from a text prompt (optionally image-to-video). Returns result URL(s). This tool waits for the result and usually returns the final URL directly. If instead it returns a generation_id with "NOT FINISHED", the job is still running: call generation_status repeatedly until it returns "Completed", then show the user the result. Never tell the user the generation merely started — finish it first. |
| ara__generate_audio | [ARA] Generate music/audio from a prompt. Returns result URL(s). This tool waits for the result and usually returns the final URL directly. If instead it returns a generation_id with "NOT FINISHED", the job is still running: call generation_status repeatedly until it returns "Completed", then show the user the result. Never tell the user the generation merely started — finish it first. |
| ara__generation_status | [ARA] Poll a generation by generation_id. Returns "Completed" with result URLs when ready, or a status to keep polling. If not ready, wait a few seconds and call this tool again — do not answer the user until it returns Completed. |
纠错与举报（发现条目失效、署名有误或涉及侵权？）
提交举报 / 纠错
侵权举报经核验成立后，我们会即时下线该条目并删除已存的内容副本。