Sentric: One answer, a tenth of a cent, no account

Solana wallet activity turned into a scored signal an agent can buy per call over x402, with the conviction weights published rather than hidden.

Solo — ingestion, scoring, payment layer, API · 2026

The problem

A trading agent does not need another dashboard. It needs one answer, fast, and it cannot sign up for anything. Every data product in this space is built for a human with a credit card: an account, a key, a monthly minimum, a rate limit tier.

That pricing model is the actual barrier. An agent that wants one signal, once, should be able to pay for exactly that and get on with it.

So the product is one call, one answer, a tenth of a cent, and no account anywhere in the path.

Architecture

Watch, score in code, sell the answer for a tenth of a cent. The caller decides what to do with it.

58 or more wallets watched continuously over a Helius websocket
Swaps on Jupiter, Raydium and Meteora captured in under 400ms
Position size 40%, hold history 20%, historical PnL 15%, rug avoidance 15%, multi wallet consensus 10%
Scores land in a store that answers by recency and conviction
x402 settles 0.001 USDC per call, with no key and no account
Dev mode bypasses payment and returns the same shapes
Sentric never trades; the caller’s agent decides

Key decisions

Payment per call, not per month

x402 turns the HTTP 402 status into a real payment handshake, so a caller settles a tenth of a cent in USDC and gets the answer. No signup, no key to rotate, no minimum. It also means the product can be metered honestly: an integration that calls once a day pays for once a day.

Conviction is five weighted inputs, in code

Position size carries 40 percent, hold history 20, historical PnL 15, rug avoidance 15 and multi wallet consensus 10. Every part is computed from on chain history, so the same trade produces the same score and a caller can argue with the weights rather than with a black box.

Latency is the product

A KOL trade is worth something for minutes. Capture runs on a websocket rather than polling, and normalization is under 400 milliseconds, because a signal that arrives after the move is just a news article.

The wallet list expands itself

A sourcing pipeline promotes wallets into the watch list rather than relying on a hand curated set that decays. A fixed list of influencers is a product that quietly gets worse every month.

It reports, it does not trade

Sentric returns what was bought, by whom, and how confident the score is. It never places an order. That boundary keeps the product honest about what it knows and keeps the operator out of a business that needs an entirely different set of controls.

Numbers

$0.001 — per API call (settled in USDC over x402)
<400ms — swap to normalized signal (websocket, not polling)
58+ — wallets watched (expanded by the sourcing pipeline)
5 — inputs to the conviction score (weights published, not inferred)
0 — accounts required (no key, no signup, no minimum)

There is no published track record of whether following high conviction signals makes money, and the case study does not claim one: that would need a backtest with slippage and a live paper trading period, and neither exists yet. Coverage is a curated slice of Solana, not the whole market, and copying a wallet is a strategy with well known failure modes that this product does nothing to fix.

sentric.sh