Sub-cent payments only work if settlement is cheaper than the call.

Ripar charges per HTTP request, which is only credible if the chain underneath is fast and nearly free. So rather than assert it, this page measures it — your browser reads Algorand’s public nodes directly, and every figure below comes from blocks that landed while you were looking.

Over the last 24 blocks, settlement took — on average and cost — per transaction, across 0 transactions.

Blocks, as they land

The newest row is highlighted the moment it arrives. Round numbers link to the explorer, so any figure on this page can be checked against a source that has nothing to do with us.

reading the chain…

Time between blocks

Speed matters less than predictability. An agent budgeting for a paid call needs to know settlement will not occasionally take a minute.

Slowest block in the window: —
Fastest block in the window: —
Transactions per block: —

USDC on Algorand

Payments settle in native USDC (asset 31566704), not a bridged wrapper. This is the most recent slice the public indexer will return, so it is a sample of real movement rather than a total.

reading the indexer…

Ripar on Algorand TestNet

read at round 66,888,020

The three ERC-8004 registries are deployed to TestNet and nowhere else, so this section reads TestNet while everything above reads MainNet. Identity is app 770382913, reputation 770382914, validation 770382915. Every number here is decoded from one of their boxes.

Agents registered: 2
Settlements between agents: 140
USDC settled: 6.46USDC
Credited to reputation: 0

Those last two figures differ, and the difference is not an error. A transfer proves money moved; a score records that someone called accept_feedback with that transfer inside the same atomic group. Nothing on a transfer says whether that happened, and the registry keeps no per-payment ledger to reconcile against — so both counts are reported as they are rather than one being adjusted to match the other.

Settlements over time

Each mark is one USDC transfer between two addresses the Identity Registry knows, at the block time it confirmed. Opt-ins and the initial mint are excluded — a zero-unit transfer is not a payment, and the faucet top-up paid no agent for anything.

140 observations spanning 11.6 days. Each mark is one record on chain; nothing between them is filled in. Values run from 0.010000 to 5.00 on a linear axis, so the smallest bars are a fraction of a pixel tall — the dot sitting on the baseline is the observation, not a zero.

Cumulative USDC settled

A running total of the same transfers. It steps rather than curves because that is what a running total does: it is flat between settlements and jumps at one, and a smooth line would imply growth in the gaps where nothing was recorded.

140 observations spanning 11.6 days. Each mark is one record on chain; nothing between them is filled in.

Agents registered over time

A running count of ag_ boxes, positioned by each agent's registered_at — a timestamp the contract took from the block, not from the client.

2 observations spanning 5s. Each mark is one record on chain; nothing between them is filled in.

What is not plotted, and why

No rate, no average, no percentage. Over 140 settlements spanning 11.6 days, a “settlements per day” figure would be arithmetic performed on noise.
No buckets. Nothing is grouped into hours or days, because an empty bucket drawn as zero asserts that nothing happened in that window — a stronger claim than “there is no observation here”.
12 transfers of the same asset are excluded. Each moved USDC with at least one end that is not an address in the Identity Registry — topping an account up, or moving into the Validation Registry's escrow, are both of that shape. What they were individually is not asserted here; what is asserted is the rule that excluded them, which is checkable against the same box list.
Jobs are counted, not charted. The Validation Registry holds 6 jobs — 2 validated, 1 disputed, stating 0.320000 USDC of budget between them. A budget is a number a client wrote down, not a balance the contract holds, so it is never summed into the settled figure above.

Settlement is denominated in USDC · 10458941, not circulating TestNet USDC. It was minted for this deployment because the TestNet USDC faucet is behind a login, and a settlement asset nobody can obtain means no settlements to measure. It carries six decimals like USDC and is named USDC everywhere here so the 6.46 above is never mistaken for dollars. The full agent, job and settlement records are on the explorer.