Proposal: a feeless settlement lane for the billing ledger
The gateway meters and bills down to micro-dollars (charged_cost per ledger row, per-user daily and
calendar-month budgets, margin per tenant × model), but it has no way to collect a single sub-cent
charge for real. That is the gap: the ledger can account for a $0.0005 request exactly, yet combining
that with a card or fiat settlement batches it into a monthly statement (or rounds it), because a card
transaction has a ~$0.30 minimum + 2.9%. So the exactness the ledger buys is thrown away at the moment of
payment.
A Nano (XNO) settlement option keeps the per-row exactness all the way to collection:
- Feeless at any amount. Nano moves value with no fee, so a micro-dollar charge (
charged_cost for
one request) can be settled on its own with no floor. The smallest price the gateway can bill is the
smallest amount it can also collect. (For scale: the Apify actor store sells results from $0.0005, and
Nano settles that exact amount with no minimum; the popular RapidAPI marketplace lists a $0.00003
minimum per call, which a card invoice under $0.50 cannot even collect.)
- Instant. Median time-to-finality under a second, so a request's charge is settled before the next
request arrives — per-call settlement instead of nightly batching.
- Self-custody, no issuer. The payer holds a Nano balance; no card network or bank can freeze or
delay a micro payment.
The seam is the ledger itself. A settlement lane would read each committed charged_cost row (or the
existing per-user budget rollover state) and issue a Nano transfer for the exact value, attributing to the
same effective user the ledger already tracks (/admin/usage/users). It stays opt-in — the gateway keeps
its current model and simply gains a per-call collection path that matches its own precision.
This is a proposal request, not a code submission; I can provide a reference implementation against the
ledger / batch-queue shape if the direction is welcome.
Proposal: a feeless settlement lane for the billing ledger
The gateway meters and bills down to micro-dollars (
charged_costper ledger row, per-user daily andcalendar-month budgets, margin per tenant × model), but it has no way to
collecta single sub-centcharge for real. That is the gap: the ledger can account for a $0.0005 request exactly, yet combining
that with a card or fiat settlement batches it into a monthly statement (or rounds it), because a card
transaction has a ~$0.30 minimum + 2.9%. So the exactness the ledger buys is thrown away at the moment of
payment.
A Nano (XNO) settlement option keeps the per-row exactness all the way to collection:
charged_costforone request) can be settled on its own with no floor. The smallest price the gateway can bill is the
smallest amount it can also collect. (For scale: the Apify actor store sells results from $0.0005, and
Nano settles that exact amount with no minimum; the popular RapidAPI marketplace lists a $0.00003
minimum per call, which a card invoice under $0.50 cannot even collect.)
request arrives — per-call settlement instead of nightly batching.
delay a micro payment.
The seam is the ledger itself. A settlement lane would read each committed
charged_costrow (or theexisting per-user budget rollover state) and issue a Nano transfer for the exact value, attributing to the
same effective user the ledger already tracks (
/admin/usage/users). It stays opt-in — the gateway keepsits current model and simply gains a per-call collection path that matches its own precision.
This is a proposal request, not a code submission; I can provide a reference implementation against the
ledger / batch-queue shape if the direction is welcome.