--- title: "The pulse" description: "The whole funnel for this x402 store, organic only: how many times a price was quoted (402s answered, not distinct agents), how many purchases settled, and how many artifacts were re-verified afterwards." canonical: "https://scvd.store/pulse" url: "https://scvd.store/pulse" --- # The pulse The whole funnel for this x402 store, organic only: how many times a price was quoted (402s answered, not distinct agents), how many purchases settled, and how many artifacts were re-verified afterwards. ## Computed at 2026-10-29T04:15:51.064Z ## Crawler correction computed at 2026-10-19T13:30:35.691Z ## Crawler correction complete true ## Crawler correction scan - **KV keys read** — 1795 - **Keys listed** — 1101727 - **List pages** — 1595 - **Cached pages** — 1503 - **Complete** — true ## House flag policy House traffic (the proprietors' own wallets and tests) is flagged at the till and excluded from every organic figure. Family doesn't make the paper. ## All time - **Organic challenges** — 289309 - **Organic payments presented** — 895 - **Organic settled** — 289 - **Total organic settled** — 297 - **MPP organic settled** — 8 - **Organic declines** — 606 - **Organic rechecks** — 1784 - **Conversion rate** — 0.000999 - **Misbooked house** — 19 ## Months ### Item 1 - **Month** — 2026-10 - **Organic challenges** — 139417 - **Organic payments presented** — 381 - **Organic settled** — 152 - **Total organic settled** — 159 - **MPP organic settled** — 7 - **Organic declines** — 229 - **Organic rechecks** — 520 - **Conversion rate** — 0.00109 - **Known machinery** — 88546 - **Known machinery by user agent** — 0 - **Known machinery by behaviour** — 88546 - **Corrected challenges** — 50871 - **Corrected conversion rate** — 0.00299 ### Item 2 - **Month** — 2026-09 - **Organic challenges** — 108312 - **Organic payments presented** — 434 - **Organic settled** — 113 - **Total organic settled** — 114 - **MPP organic settled** — 1 - **Organic declines** — 321 - **Organic rechecks** — 851 - **Conversion rate** — 0.00104 - **Known machinery** — 126139 - **Known machinery by user agent** — 1042 - **Known machinery by behaviour** — 125097 - **Corrected challenges** — 0 ### Item 3 - **Month** — 2026-08 - **Organic challenges** — 32596 - **Organic payments presented** — 64 - **Organic settled** — 23 - **Total organic settled** — 23 - **MPP organic settled** — 0 - **Organic declines** — 41 - **Organic rechecks** — 340 - **Conversion rate** — 0.000706 - **Misbooked house** — 19 - **Known machinery** — 31148 - **Known machinery by user agent** — 6837 - **Known machinery by behaviour** — 24311 - **Corrected challenges** — 1448 - **Corrected conversion rate** — 0.0159 ### Item 4 - **Month** — 2026-07 - **Organic challenges** — 8984 - **Organic payments presented** — 16 - **Organic settled** — 1 - **Total organic settled** — 1 - **MPP organic settled** — 0 - **Organic declines** — 15 - **Organic rechecks** — 73 - **Conversion rate** — 0.000111 ## Latency ### What this is How long this store's own server took to produce a 402 challenge, in buckets. SERVER-SIDE ONLY: the clock starts when the payment gate is entered and stops when the challenge is ready, so it excludes DNS, TLS, network transit and everything the buyer's own client does. A reader measuring from outside will always see a larger number than this, and the difference is the network, not us. ### Method One counter per route class per bucket per month, bumped after the response is sent so the instrument never slows the thing it measures. Percentiles are published as the INTERVAL they fall in rather than as a point value, because that is all a histogram can honestly support. THE SAMPLE COUNT IS A FLOOR, NOT A CENSUS: increments use the shared counter ledger and are mirrored to Workers KV. Earlier unpaid-door writes bypassed that ledger and used last-write-wins KV, which could lose increments under contention. Connecting the writer preserves existing counter baselines; it does not recover missing samples. Background recording failures can still leave samples uncounted. House and infrastructure traffic are counted here, unlike everywhere else in this document, because latency is a fact about our own server doing work and our own probes run the same code a stranger does — excluding them would shrink the sample without making it truer. This is the store timing itself, which is exactly the reading nobody should take on trust: the external monitor linked from this store is the one that owes you no favours. ### External monitor ### Bucket edges ms - 50 - 100 - 250 - 500 - 1000 - 2000 - 5000 ### Routes #### Challenge ##### Samples 350247 ##### Buckets - **Over** — 42 - **U00050** — 311187 - **U00100** — 13426 - **U00250** — 14095 - **U00500** — 2028 - **U01000** — 2357 - **U02000** — 3991 - **U05000** — 3121 ##### P50 ms range - 0 - 50 ##### P95 ms range - 100 - 250 ### Truncated false ## Note The x402 funnel. Legacy organic_settled and conversion fields cover x402; total_organic_settled adds corrected native MPP sales and has no shared conversion denominator. The whole funnel is reported as recorded requests and settlements. These ratios compare those counts; they do not measure unique buyers. Unclassified traffic can include unidentified crawlers. Historical lost increments remain unresolved, and subtracting machinery found in event rows from a separately counted request total does not establish an exact buyer denominator. Organic only: house traffic is the proprietors' own wallets and tests, flagged at the till and excluded here exactly as it is excluded from /stats. A conversion rate of null means no challenges were counted in that window; it does not establish that none were served. These are counts and nothing else — no user-agents, no referrers, no wallet addresses, no per-visitor rows — and the counters they read predate this endpoint, so the collection cannot have been tuned to flatter the publication. Recorded settlements alone do not establish delivery. The checkout delivers before settlement, and separate evidence is needed to verify that the buyer received the goods. Two checks look for exactly that: a delivery audit that flags a settled sale whose goods never went out, and an hourly walk of USDC arriving on Base against the certificates minted, which is independent of every write this store makes. If either ever finds one, it goes on /corrections with a date, like everything else. Settles later reclassified from organic to house are subtracted here, exactly as /stats subtracts them, and the amount taken out is published beside the figure as misbooked_house rather than left implicit. WHAT THAT CORRECTION DOES NOT REACH, said plainly because it inflates the denominator's opposite: the reclassification ledger freezes a SETTLE count per wallet and nothing else, so the challenges, declines and re-checks those same wallets generated are still counted organic here. These fields can include traffic later classified as house, while lost increments can bias the counts in the other direction. The ratio is therefore not a guaranteed upper or lower bound on buyer conversion. THE FUNNEL'S MIDDLE IS DERIVED, NOT SEPARATELY METERED: the till books every payment actually presented as exactly one of settled or declined, so organic_payments_presented is their sum and cannot drift from the two numbers published beside it. And a naming correction, dated 2026-08-27: the field once called organic_verifies is now organic_rechecks, because it counts free re-checks of already-issued artifacts at /api/verify — NOT the x402 verify step — and the old name kept being read as the protocol step it never was. ## Verify URL ## Signing key ## Statistics summary - **Computed at** — 2026-10-29T04:15:51.064Z - **Refresh** — hourly - **Stale** — false ## Published counts ### Rule Every published count ships with its instrument, its population, its window, its exclusions, and any cap or floor that bounded it, on every rendering. The rows below are that, one per number this route serves; a number without a row cannot reach this route (test/published-counts.spec.ts). ### Rows #### Item 1 ##### Path all_time.organic_challenges ##### Kind count ##### Unit 402 challenges issued ##### Instrument the payment gate, one count per 402 sent (lib/metrics.ts recordChallengeIssued) ##### Population recorded 402 challenges after named exclusions; requests, not unique buyers ##### Window since 2026-07-22, the store's first day ##### Exclusions - house traffic: the proprietors' own wallets, agents and tests, flagged at the till - infrastructure: crawlers, scanners, monitors and indexes by name (lib/channel.ts, lib/crawlers.ts) ##### Floor historical unpaid-door increments lost under contention remain absent; background recording can fail and the KV mirror can lag the shared ledger #### Item 2 ##### Path all_time.organic_payments_presented ##### Kind count ##### Unit payments presented ##### Instrument derived: organic_settled + organic_declines, never a meter of its own (services/pulse.ts) ##### Population every payment an organic buyer actually presented; each is booked as exactly one of settled or declined ##### Window since 2026-07-22, the store's first day ##### Exclusions - house traffic: the proprietors' own wallets, agents and tests, flagged at the till - infrastructure: crawlers, scanners, monitors and indexes by name (lib/channel.ts, lib/crawlers.ts) #### Item 3 ##### Path all_time.organic_settled ##### Kind count ##### Unit settled purchases ##### Instrument the till (lib/metrics.ts recordSettlementKeyed), with eligible legacy counter floors raised hourly from retained settlement records and certificates (services/counter-raise.ts); confirmed x402 purchase recovery retains missing records without replaying counter increments, with the reclassification ledger already subtracted ##### Population organic_payments_presented ##### Window since 2026-07-22, the store's first day ##### Exclusions - house traffic: the proprietors' own wallets, agents and tests, flagged at the till - family settles reclassified from organic to house after the fact (/corrections) ##### Floor KV read-add-write before 2026-09-11 could drop a count under a burst; since then the counter ledger's single writer holds the truth and KV mirrors it (services/counter-ledger.ts) #### Item 4 ##### Path all_time.total_organic_settled ##### Kind count ##### Unit settled organic purchases across x402 and native MPP ##### Instrument derived: organic_settled + mpp_organic_settled (services/pulse.ts) ##### Population the disjoint payment-protocol ledgers; sales, not HTTP successes or distinct buyers; no shared conversion denominator ##### Window since 2026-07-22, the store's first day ##### Exclusions - house traffic: the proprietors' own wallets, agents and tests, flagged at the till - family settles reclassified from organic to house after the fact (/corrections) ##### Floor KV read-add-write before 2026-09-11 could drop a count under a burst; since then the counter ledger's single writer holds the truth and KV mirrors it (services/counter-ledger.ts); native monthly mirrors may lag their durable sale records #### Item 5 ##### Path all_time.mpp_organic_settled ##### Kind count ##### Unit settled organic native MPP purchases ##### Instrument monthly native sales summaries with house corrections applied once (services/mpp-sales.ts readMppSalesMonth) ##### Population native MPP sales retained by the monthly ledger, disjoint from the legacy x402 settle counters ##### Window since 2026-07-22, the store's first day ##### Exclusions - house traffic: the proprietors' own wallets, agents and tests, flagged at the till - validated native house corrections; retries are idempotent sale records ##### Floor a sale whose accounting write failed is absent; the monthly mirror can lag the durable record #### Item 6 ##### Path all_time.misbooked_house ##### Kind count ##### Unit settles reclassified from organic to house ##### Instrument the reclassification ledger (services/reclassify.ts) ##### Population the raw organic settle count before subtraction; published so the subtraction can be checked ##### Window since 2026-07-22, the store's first day ##### Exclusions none #### Item 7 ##### Path all_time.organic_declines ##### Kind count ##### Unit payments presented and refused ##### Instrument the decline desk, one count per refused presentation (lib/declines.ts) ##### Population organic_payments_presented ##### Window since 2026-07-22, the store's first day ##### Exclusions - house traffic: the proprietors' own wallets, agents and tests, flagged at the till - infrastructure: crawlers, scanners, monitors and indexes by name (lib/channel.ts, lib/crawlers.ts) #### Item 8 ##### Path all_time.organic_rechecks ##### Kind count ##### Unit free re-verifications of already-issued artifacts at /api/verify ##### Instrument the verify route's own counter (routes/verify.ts recordVerifyCall) ##### Population every verify call; not a funnel step, a different event by different callers ##### Window since 2026-07-22, the store's first day ##### Exclusions - house traffic: the proprietors' own wallets, agents and tests, flagged at the till - infrastructure: crawlers, scanners, monitors and indexes by name (lib/channel.ts, lib/crawlers.ts) #### Item 9 ##### Path all_time.conversion_rate ##### Kind rate ##### Unit organic_settled / organic_challenges ##### Instrument derived at read (services/pulse.ts rate) ##### Population organic_challenges; null when no challenges were counted, zero when challenges were counted and no settlements were counted; not unique-buyer conversion ##### Window since 2026-07-22, the store's first day ##### Exclusions - house traffic: the proprietors' own wallets, agents and tests, flagged at the till - infrastructure: crawlers, scanners, monitors and indexes by name (lib/channel.ts, lib/crawlers.ts) #### Item 10 ##### Path all_time.known_machinery ##### Kind count ##### Unit 402 challenges re-read as machinery ##### Instrument the hourly correction walk re-reading raw rows with today's crawler table and the behavioural walk detector (services/reclassify.ts) ##### Population organic_challenges; published beside the recorded figure, never subtracted from it ##### Window since 2026-07-22, the store's first day ##### Exclusions none #### Item 11 ##### Path all_time.known_machinery_by_user_agent ##### Kind count ##### Unit 402 challenges whose client named a machine ##### Instrument the correction walk's user-agent half (services/reclassify.ts) ##### Population known_machinery ##### Window since 2026-07-22, the store's first day ##### Exclusions none #### Item 12 ##### Path all_time.known_machinery_by_behaviour ##### Kind count ##### Unit 402 challenges whose client walked the catalogue like a machine ##### Instrument the correction walk's behavioural half: WALK_MIN_ITEMS distinct doors inside WALK_WINDOW_MS (lib/walkers.ts) ##### Population known_machinery ##### Window since 2026-07-22, the store's first day ##### Exclusions none #### Item 13 ##### Path all_time.corrected_challenges ##### Kind count ##### Unit 402 challenges ##### Instrument derived: organic_challenges − known_machinery, only when the walk completed (services/pulse.ts) ##### Population organic_challenges ##### Window since 2026-07-22, the store's first day ##### Exclusions - house traffic: the proprietors' own wallets, agents and tests, flagged at the till - infrastructure: crawlers, scanners, monitors and indexes by name (lib/channel.ts, lib/crawlers.ts) - machinery the correction walk found after the fact #### Item 14 ##### Path all_time.corrected_conversion_rate ##### Kind rate ##### Unit organic_settled / corrected_challenges ##### Instrument derived at read (services/pulse.ts) ##### Population corrected_challenges; the same three states as conversion_rate ##### Window since 2026-07-22, the store's first day ##### Exclusions - house traffic: the proprietors' own wallets, agents and tests, flagged at the till - infrastructure: crawlers, scanners, monitors and indexes by name (lib/channel.ts, lib/crawlers.ts) - machinery the correction walk found after the fact #### Item 15 ##### Path months[].organic_challenges ##### Kind count ##### Unit 402 challenges issued ##### Instrument the payment gate, one count per 402 sent (lib/metrics.ts recordChallengeIssued) ##### Population recorded 402 challenges after named exclusions; requests, not unique buyers ##### Window the named month ##### Exclusions - house traffic: the proprietors' own wallets, agents and tests, flagged at the till - infrastructure: crawlers, scanners, monitors and indexes by name (lib/channel.ts, lib/crawlers.ts) ##### Floor historical unpaid-door increments lost under contention remain absent; background recording can fail and the KV mirror can lag the shared ledger #### Item 16 ##### Path months[].organic_payments_presented ##### Kind count ##### Unit payments presented ##### Instrument derived: organic_settled + organic_declines, never a meter of its own (services/pulse.ts) ##### Population every payment an organic buyer actually presented; each is booked as exactly one of settled or declined ##### Window the named month ##### Exclusions - house traffic: the proprietors' own wallets, agents and tests, flagged at the till - infrastructure: crawlers, scanners, monitors and indexes by name (lib/channel.ts, lib/crawlers.ts) #### Item 17 ##### Path months[].organic_settled ##### Kind count ##### Unit settled purchases ##### Instrument the till (lib/metrics.ts recordSettlementKeyed), with eligible legacy counter floors raised hourly from retained settlement records and certificates (services/counter-raise.ts); confirmed x402 purchase recovery retains missing records without replaying counter increments, with the reclassification ledger already subtracted ##### Population organic_payments_presented ##### Window the named month ##### Exclusions - house traffic: the proprietors' own wallets, agents and tests, flagged at the till - family settles reclassified from organic to house after the fact (/corrections) ##### Floor KV read-add-write before 2026-09-11 could drop a count under a burst; since then the counter ledger's single writer holds the truth and KV mirrors it (services/counter-ledger.ts) #### Item 18 ##### Path months[].total_organic_settled ##### Kind count ##### Unit settled organic purchases across x402 and native MPP ##### Instrument derived: organic_settled + mpp_organic_settled (services/pulse.ts) ##### Population the disjoint payment-protocol ledgers; sales, not HTTP successes or distinct buyers; no shared conversion denominator ##### Window the named month ##### Exclusions - house traffic: the proprietors' own wallets, agents and tests, flagged at the till - family settles reclassified from organic to house after the fact (/corrections) ##### Floor KV read-add-write before 2026-09-11 could drop a count under a burst; since then the counter ledger's single writer holds the truth and KV mirrors it (services/counter-ledger.ts); native monthly mirrors may lag their durable sale records #### Item 19 ##### Path months[].mpp_organic_settled ##### Kind count ##### Unit settled organic native MPP purchases ##### Instrument monthly native sales summaries with house corrections applied once (services/mpp-sales.ts readMppSalesMonth) ##### Population native MPP sales retained by the monthly ledger, disjoint from the legacy x402 settle counters ##### Window the named month ##### Exclusions - house traffic: the proprietors' own wallets, agents and tests, flagged at the till - validated native house corrections; retries are idempotent sale records ##### Floor a sale whose accounting write failed is absent; the monthly mirror can lag the durable record #### Item 20 ##### Path months[].misbooked_house ##### Kind count ##### Unit settles reclassified from organic to house ##### Instrument the reclassification ledger (services/reclassify.ts) ##### Population the raw organic settle count before subtraction; published so the subtraction can be checked ##### Window the named month ##### Exclusions none #### Item 21 ##### Path months[].organic_declines ##### Kind count ##### Unit payments presented and refused ##### Instrument the decline desk, one count per refused presentation (lib/declines.ts) ##### Population organic_payments_presented ##### Window the named month ##### Exclusions - house traffic: the proprietors' own wallets, agents and tests, flagged at the till - infrastructure: crawlers, scanners, monitors and indexes by name (lib/channel.ts, lib/crawlers.ts) #### Item 22 ##### Path months[].organic_rechecks ##### Kind count ##### Unit free re-verifications of already-issued artifacts at /api/verify ##### Instrument the verify route's own counter (routes/verify.ts recordVerifyCall) ##### Population every verify call; not a funnel step, a different event by different callers ##### Window the named month ##### Exclusions - house traffic: the proprietors' own wallets, agents and tests, flagged at the till - infrastructure: crawlers, scanners, monitors and indexes by name (lib/channel.ts, lib/crawlers.ts) #### Item 23 ##### Path months[].conversion_rate ##### Kind rate ##### Unit organic_settled / organic_challenges ##### Instrument derived at read (services/pulse.ts rate) ##### Population organic_challenges; null when no challenges were counted, zero when challenges were counted and no settlements were counted; not unique-buyer conversion ##### Window the named month ##### Exclusions - house traffic: the proprietors' own wallets, agents and tests, flagged at the till - infrastructure: crawlers, scanners, monitors and indexes by name (lib/channel.ts, lib/crawlers.ts) #### Item 24 ##### Path months[].known_machinery ##### Kind count ##### Unit 402 challenges re-read as machinery ##### Instrument the hourly correction walk re-reading raw rows with today's crawler table and the behavioural walk detector (services/reclassify.ts) ##### Population organic_challenges; published beside the recorded figure, never subtracted from it ##### Window the named month ##### Exclusions none #### Item 25 ##### Path months[].known_machinery_by_user_agent ##### Kind count ##### Unit 402 challenges whose client named a machine ##### Instrument the correction walk's user-agent half (services/reclassify.ts) ##### Population known_machinery ##### Window the named month ##### Exclusions none #### Item 26 ##### Path months[].known_machinery_by_behaviour ##### Kind count ##### Unit 402 challenges whose client walked the catalogue like a machine ##### Instrument the correction walk's behavioural half: WALK_MIN_ITEMS distinct doors inside WALK_WINDOW_MS (lib/walkers.ts) ##### Population known_machinery ##### Window the named month ##### Exclusions none #### Item 27 ##### Path months[].corrected_challenges ##### Kind count ##### Unit 402 challenges ##### Instrument derived: organic_challenges − known_machinery, only when the walk completed (services/pulse.ts) ##### Population organic_challenges ##### Window the named month ##### Exclusions - house traffic: the proprietors' own wallets, agents and tests, flagged at the till - infrastructure: crawlers, scanners, monitors and indexes by name (lib/channel.ts, lib/crawlers.ts) - machinery the correction walk found after the fact #### Item 28 ##### Path months[].corrected_conversion_rate ##### Kind rate ##### Unit organic_settled / corrected_challenges ##### Instrument derived at read (services/pulse.ts) ##### Population corrected_challenges; the same three states as conversion_rate ##### Window the named month ##### Exclusions - house traffic: the proprietors' own wallets, agents and tests, flagged at the till - infrastructure: crawlers, scanners, monitors and indexes by name (lib/channel.ts, lib/crawlers.ts) - machinery the correction walk found after the fact #### Item 29 ##### Path latency.bucket_edges_ms[] ##### Kind constant ##### Unit milliseconds ##### Instrument the histogram's bucket edges, published so the ranges can be read (services/pulse.ts) ##### Population not a count ##### Window n/a ##### Exclusions none #### Item 30 ##### Path latency.routes.*.samples ##### Kind count ##### Unit server-side timings recorded ##### Instrument one counter per route class per bucket, bumped after the response is sent (lib/metrics.ts) ##### Population every timing of this store's own server producing a 402, house and infrastructure included ##### Window the current month ##### Exclusions none ##### Floor shared ledger mirrored to KV; earlier unpaid-door writes could lose samples under contention, and background recording can still fail #### Item 31 ##### Path latency.routes.*.buckets.* ##### Kind count ##### Unit timings in the named bucket ##### Instrument the same per-bucket counters (lib/metrics.ts) ##### Population latency.routes.*.samples ##### Window the current month ##### Exclusions none ##### Floor the same historical missing samples, background-write failures and mirror lag as samples #### Item 32 ##### Path latency.routes.*.p50_ms_range[] ##### Kind range ##### Unit milliseconds, [at least, under] ##### Instrument derived from the histogram at read (services/pulse.ts) ##### Population latency.routes.*.samples; a histogram supports an interval, never a point ##### Window the current month ##### Exclusions none #### Item 33 ##### Path latency.routes.*.p95_ms_range[] ##### Kind range ##### Unit milliseconds, [at least, under] ##### Instrument derived from the histogram at read (services/pulse.ts) ##### Population latency.routes.*.samples ##### Window the current month ##### Exclusions none