MCP App Store
by KeenEthics
Inception Analytics icon

Inception Analytics MCP Server for ChatGPT

By Inception Technology SAS

Overview

Inception Analytics provides authorized Inception users with market positioning and flow analytics, Factor Lens attribution, conditional scenarios and native market views. It also explains supported Inception methods. Availability follows each user's permissions and data coverage. Prepared analysis can be displayed in saved, owner-scoped responses; users can export saved native views as PNG and explicitly retain supported datasets. An explicitly authorized support action can send a concise gap report to Inception support. Outputs are informational, conditional scenarios are not unconditional forecasts or price probabilities, and the plugin does not execute trades or analyze uploaded client portfolios.

How to connect

  1. ChatGPT: Open the Inception Analytics listing in the ChatGPT app directory and choose Connect. ChatGPT listing Inception Analytics exposes 32 tools here.

Tools

analyze_factor_lens_liquidation

ChatGPT
Knowledge analytical preparation. Prepare governed analytical facts and, when supported, one native visual candidate; this tool does not display an App. Align Factor Lens attribution with modeled systematic positioning only when the user requests their relationship: for example, whether a negative Factor Lens contribution coincided with systematic selling or how trend followers behaved during that period. Do not use this tool for a macro/residual attribution request alone or as a fallback for unavailable Live Macro. For an immediate same-chat follow-up use the exact asset/start/end of the preceding Factor Analysis result or its ready immediateFollowUpRoute; without dates use the latest five sessions. Keep total, macro, non-macro, position and percentile evidence separately dated; report flow significance only from the matched producer percentile. NEUTRAL / LOW FLOW does not support a material selling claim. The aligned sequence does not prove macro factors caused CTA trading or observed execution. For proof of observed execution, actual execution thresholds, proof or certain timing of real imminence, historical analogues or an unconditional future-price forecast, use apply_inception_evidence_guard. A named-market liquidation or short-covering risk question, including conditional imminence or the price where that pressure could emerge, is a conditional analysis from governed facts. Establish the explanation from the returned facts, then select native views only when their market, dates and scenario match the requested claim. Render each selected candidate once within render_knowledge_response, immediately after its supporting paragraph, using its returned prepared_visual_id; use a single combined view when it already covers them. Another scenario requires an explicitly qualified comparison chosen by the client. Apply visual_support for native-view relevance and selection. Missing matching native payload means text_only with the available facts, never a baseline or recalculated custom-path substitute. Do not duplicate views. Respect explicit no-chart requests and terminal or complete-approved answers. No id means no prepared visual; keep the available facts without substituting an unrelated App.

apply_inception_evidence_guard

ChatGPT
Use for proof of observed/executed liquidation or covering, actual execution thresholds, proof or certain timing of real imminence, observed historical analogues or event frequencies, unconditional next-session price forecasts, and retired client portfolio or exposure/P&L analysis. Never request client holdings or exposures; explain the withdrawal and offer market analysis. Distinguish those evidence requests from a hypothetical price at which modeled risk could change: a named-market liquidation or short-covering risk question, including conditional imminence or the price where that pressure could emerge, uses prepare_analysis_context ('conditional_prices' for prices, otherwise 'auto'), not this guard. For a combined systematic and broader positioning analysis use prepare_analysis_context with intent='positioning_intelligence'. For a positioning brief, explicit next-session unchanged-price adjustment, named High Buy/High Sell reference at the published close or conditional reactivation prices, use prepare_analysis_context with the matching intent. Prefer get_calculation_lineage for a supported definition, equation, interpretation or documentary topic, even when a market is mentioned. Select the matching metric_id from its enum and model_scope only from returned provenance; preserve explicit limitations. Use search_with_evidence once for an unsupported topic or broader documentary coverage not supplied by lineage; do not duplicate a sufficient lineage result or bypass a denial. Current sector/universe screens use the available overview; do not request an arbitrary asset or reinterpret a requested sector. Pass the current question about the unsupported Inception operation, without conversation history or unrelated content. Its refusal is terminal for that operation: preserve the supplied refusal and do not bypass it with another Inception tool or relabel other sources as Inception evidence. This does not restrict separately requested sources. Except for status 'continue' (terminal=false): that result is not a refusal, so call its route and keep its limit beside the dependent claim.

explain_factor_lens_residual

ChatGPT
Explain the approved Factor Lens Residual versus model-fit distinction. A large Return Residual is return not attributed to the selected common macro factors in that window; R-squared is window- and regime-dependent. Do not infer numeric R-squared, alpha, observed fundamentals or current values from this conceptual result. Before drafting, read all four required_reading families with read_analysis_dataset, following each next_cursor to null, including not_applicable families. The first return is preparation only; preserve all evidence, rule scopes, conditions and exceptions. Do not repeat preparation automatically or present another source as the missing Inception evidence after a failed read. Respect separately requested sources.

export_visual_image

ChatGPT
Export an existing authorized native snapshot as PNG without refresh or calculation. Supports Balance Sheet, Factor Analysis and Simulator, including saved custom price paths. Prepare a requested path with prepare_price_scenario first, then reuse its prepared_visual_id. Omit panel or use 'full' to capture the complete saved initial view with its native panel settings. Other panels/views fail explicitly. Returns MCP image content and a persistent authenticated PNG resource for embedding identical bytes in reports. Each successful export saves a new owner-scoped image; repeated exports create separate stored copies. Does not supply a missing scenario view.

get_calculation_lineage

ChatGPT
Preferred first documentary lookup for a supported metric_id: definitions, equations, units, interpretation and time conventions, plus exact approved Factor Lens topics. Topic IDs cover overview, factor_set, input_profile, components, analysis_vs_insight, the eight individual factors, usd_residualization_hierarchy, missing_zero, residual_model_fit and r_squared_interpretation (all prefixed factor_lens_). Conceptual topics return definition_status=not_applicable and formula=null; unresolved mathematical definitions remain unresolved. Use model_scope only from returned provenance. This describes documentation and does not recompute values. If the requested metric is absent, use search_with_evidence once for the exact concept; preserve explicit evidence insufficiency.Ambiguity or an unresolved definition cannot be filled with a generic formula. D1F Lambda position and capacity are lambda-based, not sized from AUM. Calibration values remain restricted. This does not recalculate published values. analysis_method_ and analysis_presentation_ topics return the complete shared MoneyFlow analysis method and each format's presentation rules, as named by a prepare_analysis_context result's method_topics; pass that result's market.model as model_scope for these topics only.

get_factor_analysis_app_data

ChatGPT
Refresh one period inside the open Factor Analysis app.

get_flow_distribution_app_data

ChatGPT
App-only read path for the Market Overview Today/Tomorrow flow popover. It derives the Platform-style Laplace view exclusively from public MoneyFlow v2 quantiles and live-flow fields.

get_forecast_distribution_app_data

ChatGPT
App-only read path for the forecast card's empirical flow-impact popover. It is intentionally hidden from model tool selection.

get_live_app_data

ChatGPT
App-only ephemeral live read for an owned prepared view. Saved analyses remain unchanged.

get_market_coverage

ChatGPT
Discover Financial Flows and Factor Lens coverage, query symbols, quotation units, reference contract/calendar facts, dataset availability and configured positioning reports. Use no arguments for all markets, markets for exact symbols/names, or market_group for an English/French group alias; selectors intersect. universe defaults to USD; use EUR only for Factor Lens EUR coverage. Read each product's status and non-null query_asset before a data query. Preserve temporarily unavailable and excluded members. history_source_coverage reports data-field, trading-date, dated-contract and predecessor checks for up to 64 trading sessions. completeFields share the effective session count; dataFields contains remaining checks. A dated or unknown check status does not certify current data. Quotation units describe the market convention, not every historical price. Positioning reports cover configured Inception inputs only; omitted reports are not proof that external data does not exist. Contract/calendar references are indicative, not a verified trading schedule or tick-value specification. live_data describes observed upstream feed facts, not a usable price in this session. Keep delayed entitlement, subscription acknowledgement and dated contract quotes separate. meta.models usability is true exactly when the Inception symbol is a key in that model's serving asset dictionary; it is independent of feed receipt and user access. Missing model metadata means dictionary membership was not reported. Quote time, receipt time and feed check time are distinct and do not inherit the catalogue's data_as_of. feed_type is aggregate feed information; use each model's entitlement for its selected contract. This tool returns catalogue facts and gaps, not analytics, prices or proof of access. Subsequent authenticated tools check access and analytical eligibility independently. Use the returned data_as_of and check-window dates when describing freshness. Large selections have pagination.next_call; follow it with unchanged arguments. An exceptionally large record has detail_call for its customer facts; read all detail pages before interpreting it. Do not query each asset again after a group lookup. For subsequent MoneyFlow group analysis, prepare and read one member completely before the next; retain successful results when another member is unavailable.

prepare_analysis_context

ChatGPT
Knowledge analysis context: risk briefs, dated conditional P1/P2 and governed follow-ups. Prepare dated MoneyFlow facts for one named market. Choose intent='daily_report' for Daily Report or Positioning Intelligence report requests. Keep 'positioning_report' for other financial-flow reports and 'positioning_intelligence' for dated systematic/participant facts and existing technical consumers. Targeted flow and Balance Sheet questions retain their own routes; do not turn them into reports. For cross-market cohort, COT/Managed Money (MM) or Trend-component comparisons, use prepare_positioning_balance_sheet once per market and requested scope. Reuse each result and complete its declared method reads. Do not add a general analysis context when those facts and method cover the request. A changed market, period or scenario needs its own preparation; otherwise reuse the same facts and prepared handle. These three intents default to 64 historical sessions (maximum 64). Daily supplies the last ten completed sessions and their preceding close, with complete last-close flat and D2/U2 outputs. All analytical intents collect at least five completed sessions; history_sessions selects the requested retrospective window within that common context. Choose intent='brief' for a positioning risk brief; 'daily_flat' for the modeled adjustment at the next unchanged close; 'daily_zones' for named High Buy/High Sell entries and membership at the published reference close; 'conditional_prices' for the price at which liquidation, short covering or buying/selling reactivation risk would emerge. Choose 'auto' for consolidation and next-five-session flow questions. The result is a compact model view: it holds the facts once and the selected method with its format rules, names the same method_topics for optional get_calculation_lineage detail, and gives its instructions and delivery for writing, language and output requirements; apply the returned method's conditions, evidence requirements and exceptions. If preparation or a detail read fails, explain its precise limit; do not present a complete report, invent a dataset identifier or loop through preparation. Reuse the same dataset for additional detail. The user need not say WATCH or ask for two prices. For named markets or a catalogue-resolved group call once per market; do not call across an entire unrestricted universe. For a group Daily, finish preparation for each member before preparing the next. Retain valid results and use their datasets for details; a partial status alone is not a reason to prepare them again. Set requested_output_language only when the user explicitly asks for that output language; language is the conversation language, not an override. Use the result's delivery.output_language throughout the answer, including component text, with detail proportionate to the explanation. For latest or today's context omit as_of; never infer it from today's calendar date. as_of refers to the loaded reference close, not the target session. Contract and dates are resolved from the current Platform catalogue. This tool does not publish HTML/email, schedule an automation, train or approve a model. negative_lambda_requires_rnd excludes the model pending R&D to prevent negative lambdas. For this or lambda_evidence_unavailable, return only the tool's localized message for that market. No diagnosis, narrative or bypass using other tools or prior facts. When the answer relies on the named market's current allocation, today's and tomorrow's flow or Price Insight Today zones, or on an explicit comparison with the live price now, use prepare_market_overview once with assets set to these exact named markets (not systematic_history), unless a returned visuals entry already shows it, and apply its returned method. For proof of observed execution, actual execution thresholds, proof or certain timing of real imminence, historical analogues or an unconditional future-price forecast, use apply_inception_evidence_guard. A named-market liquidation or short-covering risk question, including conditional imminence…

prepare_factor_analysis

ChatGPT
Prepare historical factor return attribution, beta exposure or risk decomposition for a named target and requested window. The view parameter selects individual panels or supported combinations. The result reports the effective window and available factors. Beta describes sensitivity; return contribution and risk contribution are distinct measures. This tool prepares an analytical result without displaying an App. When a supported native view is available, the result includes a prepared_visual_id identifying that preparation. The prepared view can be displayed by render_knowledge_visual or included in render_knowledge_response. The returned source dates, coverage and availability describe the prepared result.

prepare_factor_insight

ChatGPT
Prepare the current five-session macro and non-macro return attribution, or a same-family market comparison. The result reports the selected window, available family rows and source status. Optional intraday data can be unavailable while the five-session result remains usable. Contributions describe historical model attribution rather than economic causality. This tool prepares an analytical result without displaying an App. When a supported native view is available, the result includes a prepared_visual_id identifying that preparation. The prepared view can be displayed by render_knowledge_visual or included in render_knowledge_response. The returned source dates, coverage and availability describe the prepared result.

prepare_flow_distribution

ChatGPT
Prepare a named market's historical flow distribution and the percentile of the requested flow observation. The result identifies the flow metric, horizon and available reference sample. A historical percentile describes the observation's position in that sample; it is not an event probability. This tool prepares an analytical result without displaying an App. When a supported native view is available, the result includes a prepared_visual_id identifying that preparation. The prepared view can be displayed by render_knowledge_visual or included in render_knowledge_response. The returned source dates, coverage and availability describe the prepared result.

prepare_market_overview

ChatGPT
Prepare an overview, comparison or supported ranking for the requested market symbols. Available measures include positioning, allocation, price insight, separate downside and upside stress, quotes and modeled flows. Historical systematic positioning supports explicit dates, calendar-month windows or completed-session counts; these window types have different meanings. The result reports coverage and unavailable measures. This tool prepares an analytical result without displaying an App. When a supported native view is available, the result includes a prepared_visual_id identifying that preparation. The prepared view can be displayed by render_knowledge_visual or included in render_knowledge_response. The returned source dates, coverage and availability describe the prepared result.

prepare_positioning_balance_sheet

ChatGPT
Prepare a named market's positioning balance sheet, including available COT, Managed Money and participant-cohort data. The result identifies publication dates, contracts and supported classifications. Regulatory classifications and publication schedules can differ between venues. This tool prepares an analytical result without displaying an App. When a supported native view is available, the result includes a prepared_visual_id identifying that preparation. The prepared view can be displayed by render_knowledge_visual or included in render_knowledge_response. The returned source dates, coverage and availability describe the prepared result.

prepare_positioning_paths

ChatGPT
Prepare a named market's historical systematic positioning or standard conditional five-session positioning paths. The result distinguishes historical positions from conditional future trajectories and reports the applicable dates and contracts. These paths describe positioning under stated conditions, not forecasts of market prices or observed executions. This tool prepares an analytical result without displaying an App. When a supported native view is available, the result includes a prepared_visual_id identifying that preparation. The prepared view can be displayed by render_knowledge_visual or included in render_knowledge_response. The returned source dates, coverage and availability describe the prepared result.

prepare_price_scenario

ChatGPT
Prepare a conditional price-path or shock analysis for a named market using the supplied scenario arguments. Without scenario inputs, supported results include saturation, raw-signal sensitivity or a saturation exit boundary. The result identifies the applied price reference, active contract, dates and scenario. Modeled responses describe the stated conditions rather than observed orders. This tool prepares an analytical result without displaying an App. When a supported native view is available, the result includes a prepared_visual_id identifying that preparation. The prepared view can be displayed by render_knowledge_visual or included in render_knowledge_response. The returned source dates, coverage and availability describe the prepared result.

prepare_price_to_flow_sensitivity

ChatGPT
Prepare a named market's published daily price-to-flow response around its reference close. The result contains supported price points, modeled flow responses, contract and session dates. It describes conditional sensitivity to nearby price changes, rather than observed executions, a historical flow distribution or a five-session positioning scenario. This tool prepares an analytical result without displaying an App. When a supported native view is available, the result includes a prepared_visual_id identifying that preparation. The prepared view can be displayed by render_knowledge_visual or included in render_knowledge_response. The returned source dates, coverage and availability describe the prepared result.

query_factor_lens

ChatGPT
Read governed Factor Lens data: mode='discover' lists exact fields and assets; live_fields additionally exposes live_macro_attribution and intraday_macro_attribution from governed USD exhibits, separately from historical typed fields. No date controls or EUR substitution for these live fields; preserve source freshness limitations. 'inline' returns up to 20 asset/metric pairs; 'file' preserves the complete requested observations when file mode is available in the input schema. Use period_months for calendar months, paired start/end for an exact window, or all_available alone for stored history. Preserve effective dates and field units: signed absolute return contributions, unsigned relative magnitudes; beta and risk contributions are distinct. Use visual mode only when available, with structured intent for a requested presentation not covered by a native Factor App; never pass the raw question as intent. Market groups default to USD unless EUR is requested. These factor measurements do not establish observed executions or economic causality. Read file results through read_analysis_dataset using their dataset_id. Current family monitoring uses prepare_factor_insight; reuse its ready result without reconstructing it here. Prefer get_calculation_lineage for a supported definition, equation, interpretation or documentary topic, even when a market is mentioned. Select the matching metric_id from its enum and model_scope only from returned provenance; preserve explicit limitations. Use search_with_evidence once for an unsupported topic or broader documentary coverage not supplied by lineage; do not duplicate a sufficient lineage result or bypass a denial. For a sector or group, first call get_market_coverage with the requested market_group. The response includes both Financial Flows/Systematic Trend and Factor Lens coverage states, including unavailable members. Use the relevant product's non-null query_asset only for covered markets; retain other members and their reasons in the coverage statement. Never substitute the other product's membership, the navigation list or an invented subset. For proof of observed execution, actual execution thresholds, proof or certain timing of real imminence, historical analogues or an unconditional future-price forecast, use apply_inception_evidence_guard. A named-market liquidation or short-covering risk question, including conditional imminence or the price where that pressure could emerge, is a conditional analysis from governed facts.

query_live_data

ChatGPT
Read current Inception data, not an App. Choose exactly one domain: money_flow (live_price, live_flow_lots) or factor_lens (live_macro_attribution, intraday_macro_attribution). Discover first for contracts. Inline returns bounded observations; file, where supported by the selected domain/profile, creates an immutable owner-scoped snapshot with authorized reading handles. A new call acquires a new snapshot, but never makes a stale upstream quote fresh. Report source time separately from retrieval time, delayed/closed/stale/missing states, contract, units and model basis. Conditional CTA flows are estimates, not observed orders. A signed Laplace score (-100..100) is not an empirical rarity percentile (0..100). Never invent missing percentiles or intensity. Factor Lens attribution is not CTA flow. No historical dates, custom scenarios or cross-domain fallback are accepted.

query_money_flow

ChatGPT
For a conditional price-path answer where a native plot helps, use prepare_price_scenario with scenario_intent='custom_path' and the user's exact scenario inputs; it returns both native facts and a visual candidate. Do not query raw Simulator snapshots first solely to recreate a chart. Read governed Financial Flow data. Current quotes use live_price; conditional today/tomorrow estimates use live_flow_lots. Preserve meta.live_data and signed daily score basis, not an empirical rarity percentile. Use mode='discover' for exact fields and assets, 'inline' for small lookups, or 'file' for complete requested series and exports; full history uses mode='file', limit='max'. At most 20 asset/metric pairs per call. Use one asset with price_list for an explicit custom Simulator path. Preserve contracts, dates, units and missing-data reasons; an export contains only requested metrics. Do not present remembered values or public data as Inception values. Read file results through read_analysis_dataset using their dataset_id. For a combined systematic and broader positioning analysis use prepare_analysis_context with intent='positioning_intelligence'. For a positioning brief, explicit next-session unchanged-price adjustment, named High Buy/High Sell reference at the published close or conditional reactivation prices, use prepare_analysis_context with the matching intent. For a sector or group, first call get_market_coverage with the requested market_group. The response includes both Financial Flows/Systematic Trend and Factor Lens coverage states, including unavailable members. Use the relevant product's non-null query_asset only for covered markets; retain other members and their reasons in the coverage statement. Never substitute the other product's membership, the navigation list or an invented subset. For proof of observed execution, actual execution thresholds, proof or certain timing of real imminence, historical analogues or an unconditional future-price forecast, use apply_inception_evidence_guard. A named-market liquidation or short-covering risk question, including conditional imminence or the price where that pressure could emerge, is a conditional analysis from governed facts. In discover omit assets, market_group, lookback and price_list.

read_analysis_dataset

ChatGPT
Read sections and cursor pages from an existing dataset_id or prepared visual identifier. Use a declared read.section directly on the same dataset and compatible filters. Use section='manifest' to discover unknown section and column names; follow its next_cursor until the required section is found, then stop discovery. Copy exact names; read.field identifies nested row content, not a section or selected column. Only exhaustive discovery requires every manifest page. For an export, complete every page of its required sections. Dates (start/end) and columns filter data sections only; omit these filters on manifest and native export sections. Keep the dataset, section and filters consistent across next_cursor pages. A section is incomplete while next_cursor remains. Complete analytical delivery sections without column selection. Analytical and documentary pages contain at most 200 rows and 48,000 serialized characters for the complete CallToolResult envelope, not bytes or tokens. Required evidence and completion rules are supplied by the preparation/page reading_strategy, required_reading and transport_contract. With explicit start and end on data.systematic_facts.history.sessions, sequenceSummary aggregates the full stored close-to-close window before display rounding, independently of pages and selected columns; it does not recompute the model. Gross sales are positive and net flow is signed. An unavailable summary preserves readable rows and cursor. When declared, native_archive exports exact stored native JSON as gzip/base64 chunks; native_renderer exports the packaged HTML renderer without market data. Each native export response is limited to 20,000 bytes; complete its pages using its reconstruction metadata. This tool does not request fresh market data, recalculate models or render visuals. Missing or expired identifiers do not trigger replacement data.

read_knowledge_response

ChatGPT
Application-only historical response reader. Authorize before displaying saved text. Omit block_id for the manifest; provide it to load that exact native block. No preparation or recalculation.

render_knowledge_response

ChatGPT
This tool saves an owner-scoped analytical response from ordered headings, paragraphs, lists, tables and prepared native visual blocks. At least one relevant visual block is required. Visual blocks reference prepared_visual_id values from authorized preparations for the requested markets, dates and scenarios, with an optional supported native_panel. Text blocks cite an existing source_id and exactly one fact_id or pointer. Native preparation pointers address /summary/ or /payload/; canonical analytical or documentary sources support their exact paths or the explicit /data/ prefix. Block IDs are optional and can be assigned by the server. The tool validates source references and preserves submitted block order. It reads existing facts and prepared views without calculating, refreshing market data or generating charts. Saving creates an immutable response; identical submissions reuse it. Historical reading remains subject to current source access rights. Invalid source references return invalid_input with invalid_evidence_reference details; access denials, expired sources and storage failures return their corresponding errors.

render_knowledge_visual

ChatGPT
Display one prepared native analytical visual using the exact prepared_visual_id returned by an authorized preparation. The identifier binds its original facts, market, dates, conditions and native presentation. Optional commentary contains a title, introduction, evidence_refs and an optional conclusion grounded in those facts. Supplying commentary saves an owner-scoped presentation until the preparation expires; identical source and commentary reuse it. The returned presentation ID supports reloads and transport segments without new commentary. transport_chunk_index selects a segment of that same prepared result. The tool reads existing preparations; it does not refresh market data, calculate scenarios or accept chart specifications. Expired, inaccessible or mismatched identifiers return an error.

report_knowledge_gap

ChatGPT
Report an established capability gap or tool failure only after the user explicitly authorizes this write to Inception support. It creates a support record and can send an email that cannot be recalled. Repeated calls may create another report; do not automatically retry an uncertain result. Send a concise summary and relevant failure category; keep include_user_message false unless the user explicitly consents. Never report merely to bypass a scope or evidence limit.

retain_analysis_dataset

ChatGPT
Archive write only after the user explicitly requests retaining an explored selection for a later follow-up. Use the exact callback Candidate and its published parent binding. Same request_id and binding reuse the original archive; changes are rejected. Fresh source authorization remains required. Ordinary pages use read_analysis_dataset.

run_simulator_scenario_app_data

ChatGPT
App-only Simulator scenario calculation. Accepts one absolute or percentage path representation and returns updated exhibit data to the already mounted app without creating another UI resource.

screen_client_alerts

ChatGPT
CLIENT ALERT SCREENING ROUTE: for explicit alert-candidate screening, whether a commodity move or factor change is normal versus exceptional, what information is useful to send a client, whether a beta has changed unusually, call screen_client_alerts. For a macro explanation use Factor Lens; for conditional liquidation analysis use analyze_factor_lens_liquidation. A request to watch a market requires a separately configured monitor; screening does not create one. Use the returned notify/watch classifications and client-ready delivery drafts. It also reads the official Federal Reserve calendar for upcoming speeches, testimony, FOMC meetings, Board meetings, and Beige Book dates. Calendar timing is not evidence of market impact or direction. The detector is read-only: never claim that an email was sent or an alert subscription was saved unless a separate outbound delivery system confirms it. Non-Fed scheduled-event risk still requires a separately governed calendar or news source and must not be invented from Factor Lens.

search_with_evidence

ChatGPT
Use get_calculation_lineage first for a supported metric or documentary topic. Use this once for a topic outside that enum or broader documentary coverage not supplied by lineage. Do not repeat a sufficient lineage lookup or bypass an access denial. Retrieve approved definitions, methodology, limitations and visual-reading concepts. Pass only the relevant current Inception methodology question as query; omit conversation history and unrelated content. No asset is required; do not invent one. Current values and analytical windows need their governed analytical tool. Questions about contract selection or horizon also use this methodology route; the Platform selects the active contract and supplies the applicable dates. Keep retrieved definitions distinct from any scoped editorial context, historical examples and open points. A rule governing brief emphasis does not change a metric definition or suppress a fact explicitly requested. Explain only from the returned evidence; identify insufficiency without inventing a rule or broadening into unrelated searches. For a sector or group, first call get_market_coverage with the requested market_group. The response includes both Financial Flows/Systematic Trend and Factor Lens coverage states, including unavailable members. Use the relevant product's non-null query_asset only for covered markets; retain other members and their reasons in the coverage statement. Never substitute the other product's membership, the navigation list or an invented subset. For a combined systematic and broader positioning analysis use prepare_analysis_context with intent='positioning_intelligence'. For a positioning brief, explicit next-session unchanged-price adjustment, named High Buy/High Sell reference at the published close or conditional reactivation prices, use prepare_analysis_context with the matching intent. For proof of observed execution, actual execution thresholds, proof or certain timing of real imminence, historical analogues or an unconditional future-price forecast, use apply_inception_evidence_guard. A named-market liquidation or short-covering risk question, including conditional imminence or the price where that pressure could emerge, is a conditional analysis from governed facts. Before drafting, read all four required_reading families with read_analysis_dataset, following each next_cursor to null, including not_applicable families. The first return is preparation only; preserve all evidence, rule scopes, conditions and exceptions. Do not repeat preparation automatically or present another source as the missing Inception evidence after a failed read. Respect separately requested sources.

semantic_health

ChatGPT
Use only for an explicit authentication, release, Knowledge availability or service-health diagnostic. Never use health as analytical evidence or fallback for a missing market fact.

Capabilities

WritesInteractive

Example Prompts

Click any prompt to copy it.

Similar Apps

Listed on MCP App Store

Listed on MCP App Store

Add this badge to your site or README. It links to this page.

HTML

<a href="https://mcp-app-store.keenethics-labs.com/app/inception-analytics"><img src="https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg" alt="Listed on MCP App Store" width="220" height="44"></a>

Markdown

[![Listed on MCP App Store](https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg)](https://mcp-app-store.keenethics-labs.com/app/inception-analytics)