# Query Builder Tool Reference ## 55 tools, with every value their parameters accept ### The tool register [Corrected Sep 24, 2026](https://scrutica.com/corrections#query-tools-trade-and-licensing-2026-09-24) PrintScreenshot Filter the register All **55** tools. `think` Use this tool FIRST to plan your query strategy before calling any data tool. Reason about which tools to call, in what order, and why. Identify if tools can run in parallel (no dependencies) or must be sequential (one tool needs another's output). requires reasoning, plan Free-form: `reasoning` string, `plan` array `search_facilities` Search and filter compute infrastructure facilities (data centers, semiconductor fabs, packaging plants, HPC centers). Returns facility details including power capacity, GPU count, status, location, and roles\[\]: each organisation the record names, under the relationship its own source establishes (site owner, site operator, AI hardware owner, or another relationship such as PeeringDB listing holder or developer). Describe an organisation as a site's owner or operator ONLY where roles\[\] says so; withheld\_roles\[\] lists recorded attributions the cited source does not establish. Each figure is the record's own: power\_not\_operating\_reason is null when the power figure is operating capacity and otherwise names why not (announced, under construction, a campus build-out figure on an operating site, and so on), so never add these figures together as capacity. facility\_type `hyperscale_dc` · `colocation` · `edge` · `logic_fab` · `memory_fab` · `packaging` · `hpc_center` · `ai_training` · `unknown` status `announced` · `permitted` · `under_construction` · `operational` · `expanding` · `decommissioned` sort\_by `power_capacity_mw` · `total_gpu_count` · `peak_pflops_fp16` · `name` Free-form: `query` string, `country` string, `owner_org_ids` array, `owner_org_name` string, `min_power_mw` number, `max_power_mw` number, `min_gpu_count` number, `limit` number `get_facility_details` Get complete details for a specific facility including roles\[\] (each organisation under the relationship its source establishes) and withheld\_roles\[\], all relationships, investments, nearby facilities, supply chain connections through the recorded organisation, and compute capacity timeline. Use when you need the full picture of a single facility. requires facility\_id Free-form: `facility_id` string `get_nearby_facilities` Find facilities within a radius of a geographic point. Useful for geographic clustering analysis or finding what infrastructure exists near a specific facility. requires lat, lng Free-form: `lat` number, `lng` number, `radius_km` number `search_organizations` Search public organization records by name or alias, or browse by organization type. Text search returns a bounded, ranked and deduplicated set; its total matching population is unknown even when no rows are returned. Type-only browsing counts matching public organization rows before the limit; those rows are not deduplicated organizations. Country filters apply within the selected mode. result\_scope distinguishes these populations and reports actual returned rows; disclose unknown completeness, known truncation and unavailable selected records. Requires a nonempty query or org\_type. org\_type `chip_designer` · `fab_operator` · `equipment_maker` · `cloud_provider` · `ai_lab` · `colocation_provider` · `government` · `investor` · `epc` · `power_provider` · `other` Free-form: `query` string, `country_hq` string, `limit` number `get_org_profile` Get a complete organization profile: financials (revenue, capex, market cap), the facilities whose records name the organisation in any role (count of every status; operating power and operating accelerators, with figures on non-operating records listed by status and never added in; each listed facility gives only the roles its record establishes for the organisation, so a listed site is not necessarily owned or run by it), supply chain summary (top suppliers, top customers), and investment history. For deep dives on a single company. Free-form: `org_id` string, `org_name` string `get_league_table` Get ranked list of organizations by a metric. For competitive landscape analysis and identifying the biggest players. Power and accelerator ranks use operating figures only; each row lists the figures on its non-operating records by status beside them. requires rank\_by rank\_by `facility_count` · `operating_gpu_count` · `operating_power_mw` org\_type `chip_designer` · `fab_operator` · `equipment_maker` · `cloud_provider` · `ai_lab` · `colocation_provider` · `government` · `investor` · `epc` · `power_provider` · `other` Free-form: `limit` number `get_org_supply_chain` Get upstream suppliers and/or downstream customers for an organization. Returns supply chain links with value and, only where the record's own cited source states them for the pair, product/service, category, criticality score (1-10), sole-source status and alternative suppliers; attribute\_withholding gives the reason for each withheld field. direction `upstream` · `downstream` · `both` category `semiconductor_equipment` · `foundry_services` · `advanced_packaging` · `osat_services` · `memory` · `substrates` · `gpu_accelerator` · `networking` · `power_delivery` · `construction` · `cooling` · `raw_materials` · `software_ip` Free-form: `org_id` string, `org_name` string `get_chokepoint_analysis` Identify supply chain chokepoints: suppliers carrying a sole-source flag that the record's own cited source states, with their categories and any criticality the source states. An empty result means no cited source states a sole-source relationship, not that no chokepoint exists: stored flags that rest on market share, industry structure or no source are withheld. Use for "which categories rest on a source-stated sole supplier?". Free-form: `category` string, `min_criticality` number `get_country_interdependence` Analyze a country's dependency on foreign suppliers in the AI compute supply chain. Returns total links, foreign dependency percentage, sole-source foreign dependencies (only flags the record's own cited source states), and critical foreign suppliers by category. requires country Free-form: `country` string `get_full_supply_chain_graph` Get the complete supply chain graph with all edges. Returns supplier/customer org details, value, and, only where the record's own cited source states them, product/service, category, criticality and sole-source status; attribute\_withholding gives the reason for each withheld field. Use sparingly; it returns a large dataset. Free-form: `category` string `search_investments` Search investments in AI compute infrastructure: CHIPS Act grants, sovereign AI programs, hyperscaler capex, VC/PE rounds, loans. Open records return amounts, dates and programs; licensed records return an amount band, a year and a round class instead, never the amount, date or round label. Ranked largest first by band; records from partial extracts are excluded from the ranking and the count, and counted in metadata. investment\_type `chips_act_grant` · `eu_chips_act` · `sovereign_ai_program` · `hyperscaler_capex` · `vc_round` · `pe_investment` · `corporate_balance_sheet` · `project_finance` · `government_loan` · `tax_incentive` status `announced` · `committed` · `disbursed` · `cancelled` Free-form: `country` string, `org_name` string, `min_amount_usd` number, `limit` number `get_investment_timeline` Get yearly investment aggregations: record counts and the type mix per announced year, and a USD total where at least 5 licensed amounts from at least 3 recipients back it (rounded to two significant figures; exact for years holding only open records; otherwise null with the reason). Records from partial extracts are excluded and counted separately. Free-form: `year` number `get_sovereign_programs` Get sovereign AI investment records with program names, investment types, status, recorded USD amounts, announcement/commitment/disbursement dates and estimate flags. The response reports the number of returned investment records. Free-form: `country` string `get_bis_designations` Get BIS Entity List entries and other export control designations. Returns entity names, countries, designation dates, grounds for listing, and Federal Register citations. Rows may have data\_quality\_flags (provenance and caveat notes, such as a country derived from the entity address rather than named in the notice, the sourcing of a removal date, or — on rows covered only by inference from BIS's 50% Affiliates Rule — that the rule is suspended and that the ownership evidence behind the inference no longer reproduces from current data) and a coverage\_note stating the 50%-Affiliates-Rule suspension status. Results are ordered most-recent designation\_date first and capped (default 50 rows, max 200 via limit); count reports the full matching total and metadata.note flags truncation. For recent-changes questions pass designated\_after rather than reading the whole list. Free-form: `country` string, `search` string, `active_only` boolean, `designated_after` string, `designated_before` string, `limit` number `get_bis_cross_references` Cross-reference BIS Entity List designations against organizations in the Scrutica database. Match methods (match\_method field): 4-layer name matching (exact name, alias, Jaccard similarity, trigram fuzzy) plus affiliate-graph closure ("closure" — a corporate ownership/affiliate-graph walk from the designated entity; each returned row names its match\_method) and a small curated override set. Returns matched organizations with facility count, supply chain connections, match\_method, and match confidence score. Free-form: `country` string, `min_confidence` number `get_entity_list_changes` What CHANGED in the BIS Entity List — the change log behind /export-controls/changes and its RSS feed. PREFER this over get\_bis\_designations for freshness questions ('what changed this month', 'latest additions', 'recent removals'): designation rows arrive already grouped into per-Federal-Register-notice change events — one event per printed BIS rule (pinpoint-page citations of one rule merge via the federalregister.gov document-number map), each with its canonical FR citation and link, the notice title and publication date where resolved, a derived event date labeled with its source column, entity/addition/removal counts, per-country counts, a bounded entity-name sample, and the entities cross-referenced to compute-infrastructure organizations. Removal actions are separate date-keyed events (a screening list maintained from additions alone over-blocks forever). Also returns a last-N-ISO-week activity rollup counting BOTH additions and removals — a zero-count week means the Federal Register carried no BIS action that week, not that the week is missing from the rollup. Results are bounded summaries: at most 48 events per call (default 12, newest first) with truncation declared in metadata; cite totals from the count field and per-event entity\_count, never by summing truncated lists. Every underlying designation row is anchored to a Federal Register notice (Tier 1); the event date is derived (modal designation\_date, falling back to modal effective\_date) and labeled with its source. Free-form: `limit` number, `weeks` number, `company_id` string `get_country_analytics` Operating MW, accelerators and peak FP16 PFLOP/s by host country across publicly visible facility records: each figure sums operational records only, and lists the figures on records under construction, announced, retired, without a status, campus build-out figures on operating sites and earlier phases of operating clusters by status beside it, never added in. facility\_count and status\_counts cover every status. Operating power and PFLOPS are not deduplicated by physical campus; accelerators are counted once across adjudicated shared fleets. Stored scalar sums may contain estimated inputs; this operation does not infer compute from an inventory and does not establish a national inventory. The provenance estimated\_fraction counts original flagged facility records across all country facility records, including those missing a metric; it does not incorporate correction-specific estimate flags or measure the estimated share of any scalar total. result\_scope reports matching country rows before the limit, returned rows and truncation. Omit limit for all matching country rows; disclose any truncation and restrict absence claims to returned rows. Country-row completeness does not establish complete facility records or metric values. sort\_by `facility_count` · `operating_power_mw` · `operating_gpu_count` · `operating_peak_pflops` Free-form: `country` string, `limit` number `get_compute_overview` Get all-source aggregate record counts and operating facility MW, accelerators and PFLOP/s, with the figures on non-operating records listed by status beside each and status counts over every record. Includes records unavailable to the public country-analytics reader; its totals need not equal those country rows. Operating power is not deduplicated by physical campus. Failed organization/investment counts are null. `get_facility_type_breakdown` Get count and operating capacity by facility type (hyperscale DC, colocation, logic fab, memory fab, packaging, HPC, etc.); counts cover every status and each capacity figure lists the figures on non-operating records by status beside it. For understanding the composition of global compute infrastructure. `compare_entities` Side-by-side comparison of facilities, organizations, or countries. Returns a structured comparison with all relevant metrics aligned. Organization and country capacity figures are operating figures with non-operating figures listed by status; facility rows carry each record's own figures and status. For direct comparison queries. requires entity\_type, entity\_names entity\_type `facility` · `organization` · `country` Free-form: `entity_names` array `search_hardware` Search the hardware catalog for GPU, TPU, ASIC, and interconnect specifications. Returns compute performance, memory, TDP, process node, and export control status. For comparing chip specs or identifying which hardware is export-controlled. Half-precision throughput comes back as TWO distinct fields measuring different things — fp16\_tflops (dense tensor-core) and fp16\_vector\_tflops (dense non-tensor vector), which differ by up to 58x on the same part — so name which one you are quoting and never rank across them. Results are ordered by tensor throughput first, then by vector throughput among parts with no published tensor figure. category `gpu` · `cpu` · `tpu` · `asic` · `fpga` · `interconnect` · `memory` Free-form: `search` string, `manufacturer` string, `export_controlled` boolean, `limit` number `search_deployments` Search compute deployments: GPU clusters, training runs, cloud regions. Returns hardware type, GPU count, cluster name, training FLOPS, organization, and linked facility. For understanding who is deploying what compute where. deployment\_type `training_run` · `gpu_cluster` · `cloud_region` · `inference_fleet` status `announced` · `active` · `completed` · `cancelled` Free-form: `org_name` string, `hardware_type` string, `min_gpu_count` number, `limit` number `get_compute_pricing` Get cloud GPU pricing data by provider, hardware type, region, and pricing model. Returns price per GPU-hour for comparison across cloud providers. pricing\_model `on_demand` · `reserved_1yr` · `reserved_3yr` · `spot` Free-form: `provider` string, `hardware_type` string `search_trade_flows` Query recorded bilateral semiconductor trade flows (HS 8542 integrated circuits) from six sources: UN Comtrade, CEPII BACI, Taiwan Customs, China GACC, Japan e-Stat and the US Census Bureau. The sources measure the same trade (Korea -> China appears in both BACI and Comtrade), so every aggregate group is one data\_source and one direction: compare groups across sources, never add them. China GACC files each partner's imports and exports as one direction "both" (two-way trade), which an export or import query does not return. result\_scope distinguishes rows retrieved by the filtered paginated read from groups formed from those rows; matched\_count is not a database snapshot count or a count of distinct economic flows. Output can be truncated in raw and aggregate modes. Group scope records input-row count separately. Use this for recorded trade by year, partner, HS code or direction. direction `export` · `import` · `both` aggregate\_by `year` · `partner` · `hs_code` · `direction` · `none` Free-form: `reporter_country` string, `partner_country` string, `hs_code` string, `year_start` number, `year_end` number, `data_source` string, `limit` number `search_pjm_queue` Query recorded PJM interconnection queue rows. Aggregate totals are available only when a bounded read proves it retrieved every matching row; otherwise aggregation fails and filters must be narrowed. Raw results may be partial or have unknown coverage and must not be summed as complete totals. result\_scope distinguishes queue rows from groups and records the aggregate input-row count separately. Queue capacity is not delivered power or data-center capacity. queue\_status `Active` · `In Service` · `Under Construction` · `Engineering and Procurement` · `Withdrawn` · `Suspended` aggregate\_by `state` · `queue_status` · `fuel_type` · `transmission_owner` · `none` Free-form: `state` string, `is_dc_region` boolean, `min_capacity_mw` number, `limit` number `get_capacity_snapshots` Get recorded capacity snapshot rows under facility, organization and date filters, ordered by snapshot date ascending. The capped result contains the earliest matching dates; selection among equal dates is unspecified. result\_scope counts snapshot rows, which may repeat a facility. An unknown total or truncated result cannot establish completeness, latest capacity or a complete history. Use returned\_count for returned rows and disclose matched\_count and truncation when known. Free-form: `facility_id` string, `org_id` string, `year_start` number, `year_end` number, `limit` number `get_threshold_compliance` Compare the curated model training-compute estimates with numerical components of regulatory instruments. Pass `model_name` for a model, `regime_id` for an instrument and models above its numerical line, or neither for cohort counts. Qualitative rules such as the CAC Interim Measures have no numerical crossing or count. Results retain sources, estimate flags, operation types and legal status; they do not establish designation, applicability or compliance. Free-form: `model_name` string, `regime_id` string `list_threshold_capable_facilities` Rank retrieved candidate facility rows by the nameplate power recorded on operating facilities within the country filter of a regulatory regime; records not operating (announced, under construction, campus build-out figures on operating sites, earlier phases of operating clusters, retired, no status) are never ranked and are counted by status in metadata.held\_above\_min\_power. The starting inventory combines public rows with positive raw power or GPU count and public rows whose reviewed corrections supply positive power or GPU count; numerical correction data and organization-country lookups must succeed, then row corrections apply where present. Country selection uses the HQ country of the organisation recorded in the facility owner field (the AI-hardware owner when that field is empty; a recorded organisation, not an established owner), falling back to facility country. result\_scope counts this selected candidate array before the output cap, not all matching facilities or distinct sites; separate page reads are not a frozen database snapshot. This is a jurisdiction-filtered list, not a global physical-capacity ranking; use get\_country\_analytics and search\_facilities for global capacity questions. Use this for the structural question "which facilities have the physical compute capacity to support a model that triggers regulatory obligations under regime X" — the model-level question is `get_threshold_compliance`. Returns facilities with status, operating\_power\_mw, total\_gpu\_count, owner country, and is\_estimated flag. The LLM must reason from power capacity to FLOP plausibility itself, not assert "this facility CAN train a 10²⁵ model". Free-form: `regime_id` string, `min_power_mw` number, `limit` number `get_compute_visibility_index` Get the Compute Visibility Index — a per-country capacity-weighted distribution of the ownership-transparency tiers of each facility's linked organisation (Tier 1 public-filing UBO; Tier 2 licensed-database-sourced; Tier 3 documented-but-secondary; Tier 4 linked org named, no ancestors documented; Data\_Gap no linked org, not yet investigated), weighted by operating capacity with non-operating figures listed by status. Tier 4 is NOT a claim that the chain is concealed: just under half its capacity sits under an organization that declared to the global LEI registry why it reports no consolidating parent, including declarations that no parent exists. Pass `country` (ISO2) for one country's breakdown; omit for the global summary plus top-10 countries by capacity. The Index is the source of truth for "Compute Under Control" estimates by country. Backed by `mv_compute_visibility_by_country` and the `visibility_index_snapshots` table (snapshot citation handle: `scrutica:compute-visibility-index:YYYY-MM-DD`). Free-form: `country` string `get_compute_visibility_breakdown` Per-facility tier breakdown for a given country — every facility with its assigned visibility tier (Tier\_1 / Tier\_2 / Tier\_3 / Tier\_4 / Data\_Gap) and the tier\_reason string explaining the classification. Use after `get_compute_visibility_index` to drill into the facilities driving a country's tier mix. Operating facilities first by operating power, then every other record by its recorded figure. requires country tier\_filter `Tier_1` · `Tier_2` · `Tier_3` · `Tier_4` · `Data_Gap` Free-form: `country` string, `limit` number `walk_ownership_chain` Walk an organization's ownership chain upward toward the documented ultimate beneficial owner (UBO). Returns the ordered hops (max 10) with per-hop provenance: data\_source on the connecting edge, confidence (HIGH/MEDIUM/LOW), as\_of\_date, source\_class (public\_filing / private\_database / other), and the opacity classification on termination (terminus\_listed / terminus\_government / terminus\_self\_reference / terminus\_max\_hops / terminus\_cycle / terminus\_contested — the last meaning sources name different parents at equal confidence+tier and the walk stops without electing one). Use this for "who ultimately owns X" questions and for tracing PE-fund chains to their LP base. Citation: cite each hop's edge\_data\_source individually — the chain itself is not a single source. Free-form: `org_id` string, `org_name` string `search_procurement_records` Search the Scrutica sovereign-execution procurement records (USAspending federal contracts, simpler.grants.gov federal solicitations, EU TED tenders, Korean MSIT awards, manual curations). One row per primary record; the record\_type column distinguishes solicitation / award / subaward / modification / cancellation. Returns the per-record classifier breakdown (confidence + category + ai\_relevance + reasoning) plus the per-component authority tiers (announced / committed / disbursed / deployed\_chip / deployed\_facility / deployed\_utility) so the LLM can cite the MIN-of-inputs composite tier per the classifier methodology. Each record also carries its admission under the site's rule (counted only when rated high or medium AI relevance, with confidence at or above 0.7 or a person's review, and not superseded) and any data-quality notes on its own fields; the metadata states the rule. A solicitation's amount is an advertised budget, not spending. Earlier postings of a re-posted solicitation are left out unless include\_superseded is true. Use for queries like "DoD AI contracts in Q4 2025", "EU Tier-1 frontier-AI tenders", "FedRAMP awards classified as training\_compute", "awards over $10M with confidence above 0.9". Each record carries created\_at, the date it was added from its source; each awardee name is resolved against the organization alias table as its row is written. data\_source `usaspending` · `simpler_grants_gov` · `eu_ted` · `msit_korea` · `koneps` · `manual_curation` record\_type `solicitation` · `award` · `subaward` · `modification` · `cancellation` execution\_stage `announced` · `committed` · `disbursed` · `deployed` classifier\_category `training_compute` · `inference` · `research_grant` · `datacenter_construction` · `general_it_services` · `other` · `unclassifiable` classifier\_ai\_relevance `high` · `medium` · `low` · `none` sort\_by `total_value_usd` · `procurement_date` · `classifier_confidence` Free-form: `country` string, `awardee_name` string, `awarding_agency` string, `naics_psc_cpv_code` string, `fiscal_year` number, `date_from` string, `date_to` string, `min_value_usd` number, `min_confidence` number, `min_authority_tier` number, `human_reviewed_only` boolean, `include_superseded` boolean, `limit` number `get_executed_allocation_rollup` Per-country, per-stage rollup of sovereign procurement execution (announced → committed → disbursed → deployed). A record counts only when the classifier rated it high or medium AI relevance, with confidence at or above 0.7 or a person's review, and no later posting supersedes it. The headline totals are counted records at the committed stage or later; counted solicitations are reported apart as advertised budgets, not spending; low-rated records are disclosed as their own band and not counted; the rest are counted by reason; each source's record dates are given. Above $100M a headline is withheld (null, withheld: true) until the country's below-floor records are reviewed. Backed by a rollup table that database triggers keep in step with every write to the procurement records, with mean classifier confidence (the model's own score, not a measured accuracy) and the sovereign-wealth-fund execution\_stage\_note joined back from those records at read time. Returns the disclosure\_state classification (with\_records / amounts\_unrecorded / disclosed\_zero / announced\_only / none\_counted / withheld\_for\_review / structurally\_undisclosed / no\_ingest\_yet) so the LLM can frame structural non-disclosure as a disclosure state rather than as a data gap. Composite authority tier is computed MIN-of-inputs per the classifier methodology (a tier-1 facility-evidence claim does NOT inflate a tier-3 chip-allocation press report into a tier-1 deployed assertion). Use for queries like "Saudi sovereign AI execution by stage", "compare announced vs deployed for top sovereign programs", "Korean MSIT FY2024 disbursement". For per-record drill-down use search\_procurement\_records. Free-form: `country` string, `countries` array, `human_reviewed_only` boolean `get_regulatory_impact_prescore` Simulate the structural reach of a hypothetical BIS Entity List addition (or other designation type). Given a seed organization, computes the affiliate-graph closure (downward through SUBSIDIARY/JV/AFFILIATE edges, max 4 hops), the facilities operated/owned by entities in that closure, country distribution, sovereign-LP exposure breakdown (PIF, Mubadala, Temasek etc. with disclosed dollar amounts), and coordination-gap classifications. Returns an ImpactReport: closure size, affected facility count, total power MW + H100-equivalent GPUs, country rollups, sovereign-LP breakdown. NOT a legal interpretation — see caveats\[\] in the result. NOT a prediction of BIS behavior — structural reach only. Use this for "if BIS designated X, what would be caught" pre-decision analysis. designation\_type `BIS_ENTITY_LIST` · `SDN` · `EU_SANCTIONS` Free-form: `org_id` string, `org_name` string `simulate_cascade_scenario` Retrieve generic cascade assumptions and historical examples: inventory buffers, substitution definitions, and total graph counts. This handler DOES NOT execute a shock-specific propagation simulation, scale buffers, or identify affected nodes; the perturbation, organisation and duration are echoed as context only. Use get\_org\_supply\_chain and get\_chokepoint\_analysis for actual supplier relationships. Parameters are estimates, not forecast effects or a disruption timeline. Historical reference rows cover the Samsung Austin power outage (April 2021 report), TSMC materials-supply assessment (April 2022 call), and CoWoS capacity expansion (October 2023 call), each with its own primary source and authorityTier. These cases have not been used to calibrate the model. scenario\_id `tokyo-earthquake` · `iran-threat` · `scs-cable` · `taiwan-strait` · `abqaiq` Free-form: `perturbation` string, `affected_org_name` string, `duration_days` number, `include_substitution_paths` boolean `get_scenario_model` Return the structured scenario data for one of Scrutica's five named scenario models. (1) tokyo-earthquake — the corresponding dataset (memory and equipment locations, authored seismic and compound scenarios, supplier shares and HHI, dated factory accounts and a DDR3 2Gb Q2 2011 contract-price observation; site lists and radii are chosen inputs, with no production-geography percentages or calculated earthquake losses). (2) iran-threat — the corresponding dataset (missile/drone systems, launch sites, air-defense profiles and reported strikes, including damage to AWS facilities in the UAE and Bahrain). (3) scs-cable — the corresponding dataset (selected submarine-cable inventory, landing stations, incident-attribution status, authored rerouting scenarios and chokepoint zones; unsupported project and route records are withheld; cases specify assumed cable sets without capacity-loss, latency or cloud-region exposure estimates). (4) taiwan-strait — the corresponding dataset (TSMC fab facilities, substitution paths, Chinese chip production evidence, Japan–South Korea policy chronology and trade research). (5) abqaiq — the corresponding dataset (Saudi data centres and petroleum facilities, assigned substations and grid zones, dated incident accounts and conditional disruption scenarios; unresolved incident-to-site links are null, with no interception ratios or site-hardening classifications). Use the source URLs returned with the scenario evidence. Attribute scenario assumptions to the named scenario; never cite a source-code constant as a record. requires scenario\_id scenario\_id `tokyo-earthquake` · `iran-threat` · `scs-cable` · `taiwan-strait` · `abqaiq` Free-form: `section` string `compute_flop_threshold_compliance` For a named facility, retrieve its correction-overlaid record and return compute evidence only when every partition of its complete, reconciled, operational inventory is present capacity with a cited dense-FP16 rate for exactly one accelerator model. Every partition is returned with its own verdict (admitted, or held with a hold code, reason and documented alternatives that are never applied), plus provenance, precision and specification sources; caller-supplied MW and estimator overrides are ignored for named records. Unknown, future, incomplete, mismatched and held-partition records are withheld with a reason code, the partition verdicts and a bounded reported-inventory projection where available. An organization name alone is also withheld because it has no complete facility-inventory aggregate. Without facility\_id or org\_name, a positive power\_capacity\_mw runs a clearly labelled generic H100-equivalent scenario; its qualification says that it does not describe a named facility or establish legal compliance. Figures named sustained\_\* apply the MFU and interconnect assumptions listed with them; peak\_fp16\_pflop\_per\_sec is the unadjusted peak. regime\_id `eu-ai-act` · `us-eo-14110` · `ca-sb-1047` · `ny-raise-act` · `miri-tgt-monitored` · `miri-tgt-strict` Free-form: `facility_id` string, `org_name` string, `power_capacity_mw` number, `pue_override` number, `gpu_fraction_override` number, `mfu_override` number `derive_cost_index_per_provider` Return $/petaFLOP-day broken down by cloud provider, GPU type, region, and pricing model. Computes the normalization across AWS, Azure, GCP, Oracle, CoreWeave, Lambda from the corresponding dataset (the cost-index data layer). On-prem TCO 3yr/5yr surfaces alongside as the build-vs-buy boundary. Source per row: provider pricing-page snapshot date + Epoch ML hardware spec (dense BF16 TFLOP/s; no 2:4 sparsity). Filter via params; omit all to return the global rollup. Distinct from `get_compute_pricing` (which returns raw $/GPU-hour cells) — this returns the petaFLOP-day-normalized, throughput-aware rollup. pricing\_model `on_demand` · `reserved_1yr` · `reserved_3yr` · `spot` sort\_by `cost_per_pflop_day_asc` · `cost_per_pflop_day_desc` · `provider` · `gpu_type` Free-form: `provider` string, `gpu_type` string, `region` string, `limit` number `get_page_methodology` Return the methodology block for a Scrutica page. Use for "how does Scrutica calculate X?" — returns the actual current methodology text rendered on the page, not a paraphrase. Cite the route + the specific methodology section when summarising. requires route Free-form: `route` string `get_methodology_section` Return one labeled methodology subsection from the /methodology page. Wraps the methodology page's structured data layer + section anchors (e.g., "validation-6", "flop-estimation-bounds", "cascade-criticality-weighting", "threshold-derivation"). Distinct from `get_page_methodology` (which returns a single page's full methodology); this drills into a specific named subsection within /methodology. requires section\_id Free-form: `section_id` string `get_page_content` Return the rendered prose content displayed on a Scrutica page. Use for "what does the \[page\] currently say about \[topic\]?" — the model retrieves the rendered prose then summarises against the user's question. Distinct from `get_page_methodology` (which returns just the methodology block); this returns the displayed body content. requires route audience `technical` · `policy` · `both` Free-form: `route` string `recent_editorial_corrections` Return selected public correction and withdrawal notices, optionally filtered by date or affected route. This is not a history of development fixes or routine record updates. Entries identify the affected claim and corrected interpretation. Free-form: `limit` number, `since` string, `affected_path_filter` string `search_corrections_log` Search the selected public correction and withdrawal notices. Absence of a matching notice does not establish that a record has never changed. Routine record updates and private development history are outside this collection. requires query Free-form: `query` string, `limit` number `get_data_freshness` Return per-data-source freshness state — when each data source was last refreshed, the freshness-class TTL, and the current freshness color (fresh / aging / stale). Freshness classes, each fresh for the window shown: regulatory (7 days, e.g., BIS Federal Register feeds), market (7 days, e.g., trade flows), structural (90 days, e.g., facility ownership), narrative (180 days, e.g., historical scenarios). Use for "how fresh is the \[layer\] data?" or "what's overdue for refresh?". freshness\_class `regulatory` · `market` · `structural` · `narrative` Free-form: `data_source` string `get_authority_tier_provenance` Return the authority-tier provenance metadata for a facility record: data\_source, source\_url, is\_estimated, authority\_tier, data\_vintage, confidence\_score (where applicable). The Tier 1-4 hierarchy: Tier 1 = primary measurement / regulatory filing, Tier 2 = research database, Tier 3 = press / analyst, Tier 4 = inferred / estimated. The Citations Panel UI consumes this output to render per-claim provenance badges. requires facility\_id Free-form: `facility_id` string, `field` string `get_chokepoint_anatomy` Return the layer-by-layer anatomy of the AI compute supply-chain chokepoint structure — the "hourglass" shape (many AI labs and materials at the ends, extreme concentration in fab/packaging/equipment in the middle). Each layer has: concentration\_level (moderate / high / extreme), summary\_text, vintage, source. Use for "show me the supply-chain concentration map" or "where is the system most exposed?". Free-form: `layer_filter` string `get_concentration_metrics` Return HHI, editorial priority and governance component evidence for recorded chokepoint assessments. Governance scores and rankings are unavailable: country-map tags, docket presence and unvalidated grades do not establish legal jurisdiction or enforcement effectiveness. Category filtering applies before the result limit; top\_5 is the separate global concentration-priority cohort. Null HHI or priority is unavailable, not zero. sort\_by `hhi_desc` · `severity_desc` · `governance_leverage_desc` Free-form: `category` string, `limit` number `get_country_profile_summary` Return a country profile with facility counts, reported power and GPU totals, organisations, sovereign programme funding, supply-chain relationship counts, export-control access and Entity List designations. Accepts a country name or ISO alpha-2 code. requires country Free-form: `country` string `get_emerging_chokepoints` Return Scrutica's active emerging-chokepoint feed — supply-chain constraints currently elevated to watch / elevated / high / critical severity, with deadline-bound risks (e.g., "CoWoS capacity shortfall through Q4 2026", "EU power grid pre-emption Phase 1 begins 2026-09-01"). Each entry: id, severity, category, summary, evidence, deadline (where applicable), affected\_orgs, governance\_leverage. Use for "what supply-chain risks are active right now?" or "what's elevated to critical?". severity `critical` · `high` · `elevated` · `watch` Free-form: `category` string, `with_deadline_only` boolean, `limit` number `search_bis_license_actions` Search BIS annual country-licensing data (Bureau of Industry & Security export-license actions by destination country and ECCN), calendar years 2015-2022. Each row is one ECCN, destination and year for one action, approved or denied: country\_iso3, period\_year, eccn, eccn\_rank, eccn\_description, action\_type, action\_count, pct\_of\_world\_total, reference\_url. Rows carry no friction rate, and min\_friction\_pct and friction\_tier are not applied to them (the result says so); per-country friction summaries are on /export-controls (get\_page\_content). Use for "how many \[ECCN\] licences to \[country\] were approved or denied in \[year\]?". friction\_tier `low_friction` · `elevated_friction` · `anomalous_friction` · `unknown` Free-form: `country_iso3` string, `country_iso2` string, `year_start` number, `year_end` number, `eccn_prefix` string, `min_friction_pct` number, `limit` number `get_hbm_market_share` Return quarterly HBM (high-bandwidth memory) merchant market share by manufacturer. Manufacturers: SK Hynix, Samsung, Micron. Product tiers: HBM2E, HBM3, HBM3E. Sources: Astute Group, TrendForce, Yole Développement, manufacturer disclosures. Per-cell `is_estimated`, `authority_tier` and `vintage`. Use for "what's the current HBM3E share split?" or "how have HBM shares moved since 2024?". Where a quarter-and-manufacturer cell carries more than one source the divergence is surfaced rather than averaged away, and the canonical-share function resolves the cell by preferring a measured row over an interpolated one, then the lower authority tier, then the more recent vintage. product\_tier `HBM2E` · `HBM3` · `HBM3E` · `all` manufacturer `SK Hynix` · `Samsung` · `Micron` · `all` Free-form: `quarter` string, `canonical_only` boolean `get_eu_grid_status` Return EU electricity grid capacity, HVDC interconnects, and per-country grid signal — the records behind the sovereign-AI grid pre-emption analysis. Per-country: nameplate generation MW, current draw MW, headroom MW, hyperscale-DC announcement count, announced power MW vs grid headroom, pre-emption-risk flag. 16,564 records read from PyPSA-eur (TSO-disclosed network model) on 2026-05-09. Sources: ENTSO-E, PyPSA-eur, per-country TSO disclosures. Use for "where is the EU grid being pre-empted by hyperscale buildout?" or "what's \[country\]'s grid headroom for AI compute?". sort\_by `headroom_asc` · `announced_dc_power_desc` · `preemption_risk_desc` Free-form: `country_iso2` string, `include_interconnectors` boolean `get_recent_developments` Return the recent-developments feed — Scrutica's timeline of substantive changes across the records (tenant changes, ownership transfers, capacity milestones, regulatory actions, incidents). Each entry: id, effective\_date, title, body, data\_source, source\_url, confidence (high/medium/low), category (tenant\_change / ownership\_transfer / capacity\_milestone / regulatory / incident / other), affected\_facility\_id (where applicable). Use for "what's changed in the last 30 days?" or "show me \[country\]'s most recent compute developments". category `tenant_change` · `ownership_transfer` · `capacity_milestone` · `regulatory` · `incident` · `other` min\_confidence `high` · `medium` · `low` Free-form: `days_back` number, `facility_id` string, `limit` number `get_training_evidence` Return per-model training evidence — the multi-source per-model dataset Scrutica uses to cross-validate FLOP estimates against lab disclosures. Each model: training\_claim (what the lab said), facility\_power\_evidence (where it ran, how much MW available), chip\_availability\_evidence (what hardware available + BIS export-control status), financial\_evidence (capex / opex commitments), supply\_chain\_evidence (HBM / packaging / wafer availability), signal\_status (consistent / insufficient / unknown). Use for "is \[model\]'s claimed FLOP budget consistent with the available facility / chip / supply-chain evidence?". Distinct from `get_threshold_compliance` (which returns regime-crossing verdict); this returns the underlying evidence chain. signal\_status `consistent` · `insufficient` · `unknown` evidence\_type `facility_power` · `chip_availability` · `financial` · `supply_chain` · `all` Free-form: `model_name` string, `limit` number Reading The [query engine](https://scrutica.com/query) answers plain-language questions about the site’s records through these tools. This register is generated from the definitions it executes and includes every available tool, parameter and accepted value. An enumerated parameter accepts exactly the values listed; free-text parameters are marked separately. A question phrased with a declared value resolves to that value, while other wording either falls through to free-text search or is refused. ### Query limit The limit is **8** _data-tool_ calls per question, across every round. A question needing more than that comes back partial, with the shortfall named.