{"criteria_id":"agent-query-index/v1","product_name":"Per-Provider LLM-Usage Query Index — labelled legs, never merged","name_claim_statement":"Name claim, word by word: Per-Provider = every answer is separated by provider, never merged across providers. LLM-Usage = the legs are human-to-LLM usage data (class H1: what people ask LLMs, per provider). Query = the caller submits one topic/query and receives the per-leg sections for it. Index = a lookup index over the published cleared legs, not a search engine, not a ranking, and not a source of agent-issued search data. The name claims no agent-issued search data (class A1, published by no one), no ranking, and no cross-provider merging.","form_statement":"Form: per-call lookup only (POST /lookup). The free catalog and free /criteria are config disclosure (the leg list, labels, licences and price state) — they publish no leg data and are not a merged presentation. There is no digest, ranking, or merged publication form.","presentation":"per_leg_only","cross_provider_aggregation":"NONE — there is no response field that can carry a cross-provider aggregate, merged ranking, or composite score; a cross-provider question is answered as separate per-leg sections or NOT-LOADED per leg","forbidden_answer_fields":["composite_score","merged_ranking","ranking","cross_provider_total","combined_value","overall_rank"],"corpus":{"id":"agent-query-index-corpus/v1","sha256":"7b9228949e59824e96c15e70fc6550c3bc1777d76f4ae2e679a3a50508dbef76","pin_definition":"The corpus pin is the SHA-256 of the corpus's canonical serialization (the corpus object excluding its sha256 field), computed and verifiable by the published harness (harness.mjs)."},"price":{"units":10000,"label":"$0.01","state":"RULED","band":"$0.005-$0.01 per-call lookup (raw-facts band, per the record)","note":"One figure inside the record's band, proposed by the build lane; labelled RULED (price ruled 2026-11-03). Never reprice a published base call."},"legs":[{"leg_id":"anthropic-economic-index","provider":"Anthropic","leg_name":"Anthropic Economic Index","class":"H1","class_label":"what people ask LLMs — per provider","licence":"CC-BY","attribution":"Anthropic Economic Index — data released under CC-BY; attribution to Anthropic required.","freshness_rule":"Per release: the leg states its source release and sample window. Freshness is per leg, never a composite-wide claim.","conditions":"Republication clear under CC-BY with attribution."},{"leg_id":"openai-signals","provider":"OpenAI","leg_name":"OpenAI Signals","class":"H1","class_label":"what people ask LLMs — per provider","licence":"CC BY 4.0","attribution":"OpenAI Signals — data licensed under CC BY 4.0; attribution to OpenAI required.","freshness_rule":"Per Signals version: the leg states its version (the record's v3.0 files span Jul 2024-Sep 2026).","conditions":"Republication clear under CC BY 4.0 with attribution."},{"leg_id":"wildchat-aggregates","provider":"WildChat (Allen Institute for AI)","leg_name":"WildChat aggregates","class":"H1","class_label":"what people ask LLMs — per provider","licence":"ODC-BY","attribution":"WildChat aggregates — database rights under ODC-BY; attribution required. Aggregates computed from a corpus frozen May 2024, HF-Spaces population.","freshness_rule":"Corpus frozen May 2024, HF-Spaces population — this date travels with the leg wherever it is shown.","conditions":"Aggregate statistics only. Verbatim prompt republication stays out (ODC-By contents caveat)."},{"leg_id":"radar-web-graph","provider":"Cloudflare Radar","leg_name":"Cloudflare Radar web-graph material","class":"H1","class_label":"fetch targets, not questions","licence":"CC BY 4.0","attribution":"Cloudflare Radar — website graphs/images and limited attributed excerpts only, CC BY 4.0; attribution to Cloudflare required.","freshness_rule":"Per excerpt/graph: the leg states the excerpt date. Freshness is per leg.","conditions":"Website graphs/images + limited attributed excerpts ONLY. Radar API/download data is CC BY-NC 4.0 and BARRED — never a leg, never loaded, never served."},{"leg_id":"glama-display","provider":"Glama","leg_name":"Glama display leg","class":"H1","class_label":"per-platform tool popularity, as displayed under Glama's credit conditions","licence":"Conditional — Glama Terms of Service §15 (display/use in a product or service only)","attribution":"Glama — displayed under Glama's credit conditions: §15.3 per-page visible linked credit to Glama on every surface displaying this leg, and §15.4 a link to each record's own Glama listing.","freshness_rule":"Live counters, snapshot-dated: the leg states its snapshot date.","conditions":"Display only: §15.3 per-page linked credit + §15.4 per-record listing links on every displaying surface. Never bulk republication, never a dataset/feed/export, never a competing directory (§15.6).","glama_credit":"Data displayed from Glama (glama.ai) under its credit conditions."}],"barred_leg_aliases":["radar-api","radar_api","cloudflare-radar-api","cloudflare-radar-api-data"],"barred_sources":["Radar API / download data — CC BY-NC 4.0 — BARRED: never a leg, never loaded, never served","ShareChat — CC BY-NC 4.0 — commercial republication barred","LMSYS-Chat-1M — distribution barred (Prohibited Transfers); internal-only, not a leg"],"unclear_not_load_bearing":"No UNCLEAR source is load-bearing: Smithery, PulseMCP, Hagglyai, mcp.so, the MCP Registry compilation, package registries, Google Trends, BrowseComp/QREF, and OpenRouter (reported CC BY 4.0 but not firsthand terms-verified in the evaluation record) are NOT legs and never carry this composite.","composite_labels":["per-provider, not cross-provider","H1 legs: what people ask LLMs — per provider (never \"what agents ask\")","Radar leg: fetch targets, not questions","Glama leg: per-platform tool popularity, as displayed under Glama's credit conditions","WildChat figures: aggregates computed from a corpus frozen May 2024, HF-Spaces population","Legs are never merged; no composite score; no merged ranking"],"a1_absence_statement":"No agent-issued search-query data exists in this composite, because none is published by anyone. Class A1 — the searches agents themselves issue — is published by no provider, registry, or study: no production query text, categories, or rankings. This service states that absence; it never approximates A1 from H1 legs or from any proxy.","classes":{"H1":"human->LLM questions — what people ask LLMs, per provider","A1":"agent-issued web-search queries — DOES NOT EXIST anywhere: published by no one (the absence statement is the answer)"},"leg_grades":["ANSWERED","NO-MATCH","NOT-LOADED"],"matching_rule":"Per leg, mechanical: an entry matches when the normalised topic (trim, lower-case, collapsed whitespace) equals the entry's normalised topic or one of its normalised keywords. A loaded leg with no matching entry grades NO-MATCH — stated, never filled from another leg.","load_state_rule":"A leg whose data is not loaded grades NOT-LOADED — never stale or partial data dressed as the leg. Phase 1 ships with every leg NOT-LOADED at runtime; leg data is loaded by the main seat at release via the COMPOSITE_LEGS_DATA env var (JSON keyed by leg_id: {load_state:'LOADED', freshness, entries:[...]}). Pinned fixture extracts under build/legs/ exercise the LOADED path in tests only and are labelled FIXTURE.","no_composite_score":true,"no_merged_ranking":true,"legs_never_merged":true,"payment_includes":"Your payment includes one query-index lookup for the topic you submit: one answer section per requested leg (provider, class H1, licence, attribution, that leg's freshness, and its load state printed on every section), the matching entries from each loaded leg, the per-leg factors, and the A1 absence statement. Legs are never merged and no composite score or merged ranking is included. It does not include any agent-issued search-query data (class A1 — none is published by anyone), any Radar API data (barred, CC BY-NC), any verbatim WildChat prompts, or retention of your submitted topic (it is discarded; only its hash is kept).","retention":"Store-nothing beyond the answer: the submitted topic is discarded after the lookup; the service keeps only the topic's SHA-256 hash, the requested leg ids, aggregate counters, and the payment record.","standing_condition":"Standing condition: any version of this composite resting on a barred or unclear leg, or presented merged or unlabelled, fails clean-full-data and must not ship in that form.","tripwire":"Kill line: build, bank, and hold; judged at the fleet's periodic reviews. Tripwires: refunds over 5% of paid calls in any rolling 30 days, free use over 10x paid use, or a wrong answer from undisclosed stale/partial data = stop-serving. A version resting on a barred/unclear leg, or presented merged/unlabelled, must not ship in that form.","body_cap_bytes":262144,"free_tier":{"per_day":100,"min_interval_ms":5000,"on_exhaustion":"402"},"ledger_fields":["payer","ts","date","price_usdc_units","payment_tx","solana_fee_lamports","criteria_version","op","question_class","topic_hash","legs_requested","legs_answered","legs_not_loaded","house","cad_rate_usdcad","cad_rate_date","cad_basis","cad_fmv","cad_pending","refund"],"service":"query-index","version":"1.0.0","endpoints":{"criteria":"GET /criteria (free)","precheck":"POST /precheck (free: validation + per-leg load state + answerability disclosed before payment; no answer entries)","lookup":"POST /lookup (paid, $0.01 per call — RULED: one per-leg composite lookup, legs never merged)","metrics":"GET /metrics"}}