x402 spend tracking in 2026: what actually exists, and which one you want
GetAgentCap team · published 9/20/2026
If you are running agents that pay for things over x402 and you have gone looking for a way to watch that money, you have probably found the same scattered set of tools I did and struggled to tell them apart. They describe themselves in nearly identical language. They are not the same product. Here is the honest map, including the ones that compete with us.
The four categories
Explorers. x402scan.com and x402data.org. These index x402 activity across the ecosystem — transactions, endpoints, who is paying whom. They answer "what is happening on this protocol." They are excellent for that and are not built to answer "what is my fleet doing," because they have no concept of your fleet. If you want to understand x402 as a market, start here. If you want to reconcile your own spend, you will be filtering someone else's firehose.
Policy engines / enforcement layers. Sentinel, Locus, AgentGuard, and the various caps-and-kill-switch projects that show up on Show HN. These sit in your agent's request path and refuse payments that violate a rule — per-transaction limits, allowlists, hard caps, auto-kill. This is the strongest category if your requirement is that the spend cannot happen. It is also the category with the highest integration cost, because a thing that can block a payment is by definition a thing your agent has to route through, and you now own its uptime.
Cross-rail cost trackers. Centurian.ai and similar. They aggregate agent cost across multiple rails, including model spend and x402 spend. Broader scope than us, more surface to configure.
Read-only ledgers. That is us. AgentCap takes a public wallet address, nothing else, and shows you every USDC payment out of it on Base and Solana, attributed per agent, with a budget ceiling per wallet, a webhook when one crosses, and CSV export.
The question that actually decides it
Forget the feature grids. Answer one thing: do you need the payment blocked, or do you need to know it happened?
If the answer is blocked — a compromised agent, a customer-funds scenario, a hard compliance ceiling — buy an enforcement layer and stop reading comparisons. Nothing observational will save you, including us. We watch settled chain data; by the time we see a transfer, the money is gone. Layering alerts over a hole does not close the hole.
If the answer is know — and in every incident write-up I have read, from the $58 overnight bill to the agent that ran a café's back office $38k out against $9k in, the failure was that nobody knew for weeks — then enforcement is a large integration bill for a problem you do not have. What you have is an attribution problem.
Why read-only is a feature, not a limitation
Because we only ever take a public address, there is nothing to integrate, nothing in your request path to go down, and nothing we hold that can be drained. You can point AgentCap at a fleet you did not build, that someone else operates, that is already running in production, without touching a line of its code or asking anyone for a key. No enforcement tool can offer that, structurally — the power to block is the same power that has to be wired in.
What we will not claim
We are not the only thing here. The tools above are real, several are good, and some are better than us at what they do. What we have not found is another product whose entire job is per-agent attribution across Base and Solana with zero integration — which is why we built it, and why we will happily tell you to go buy an enforcement layer if blocking is what you came for. See the category breakdown at /compare.
Paste an address at getagentcap.com and see your own numbers in about a minute. Free up to ten wallets, $19 for thirty days past that.