Sending on the wrong network is the least forgiving mistake in this whole subject. Miscalculate a fee and you overpay slightly. Get the address or the chain wrong and the money is usually just gone — no support desk recovers it, no institution reverses it.
So this tool is deliberately modest about what it claims: what it can verify, it verifies properly; what it cannot, it says so. Plenty of address validators advertise checksum verification and in fact only compare character ranges. Here the two are labelled separately, and "checksum verified" and "format only" are genuinely different strengths of conclusion.
What it checks, what it does not, and the one question it cannot answer (expand)
What is genuinely verified
Addresses starting bc1 (bech32 and bech32m) get a full checksum computation. That format has error detection built into its design: drop a character, add one, or mistype one, and the check fails. This layer is real arithmetic, not a glance at the character set.
Related quirk the tool also enforces: these addresses do not allow mixed case. All upper or all lower is valid; a mixture is not.
What is only format-checked
For the following, the tool inspects character set and length only, with no checksum computed: EVM addresses beginning 0x, TRON addresses beginning T, legacy Bitcoin addresses beginning 1 or 3, and Base58 addresses in the Solana shape.
A pass here means the address is shaped correctly. It does not mean it was copied correctly, and it certainly does not mean it belongs to the person you intend to pay.
The question it cannot answer, which is also the important one
This is the section worth reading twice. The same 0x address is formatted identically on Ethereum, BNB Smart Chain, Polygon, Arbitrum, Optimism and every other EVM chain. The tool sees a string of hexadecimal characters and has no possible way to know which chain you should be sending on.
Only the recipient can tell you the network. Not this tool, not the shape of the address, not a guess based on what you sent last time. Assets sent on a network the destination does not support are generally unrecoverable, and no format check anywhere can prevent that.
What "checksum valid" does and does not tell you
It tells you the address was transcribed without error. It tells you nothing about ownership. An attacker's address passes a checksum exactly as well as your friend's does. A valid checksum on an address you got from a compromised source is a correctly copied wrong address.
Memos and tags are outside this check
Some networks and some recipients require an additional identifier alongside the address. This tool checks the address only. Omitting a required memo means the transfer succeeds on-chain while the recipient's system cannot attribute it to you — a different failure with a different recovery process, covered in withdrawal fees by network.
Nothing you paste leaves the page
The checking runs entirely in your browser. Nothing you paste is sent anywhere, recorded or stored, and it is gone when you reload the page.
The step to take regardless of what this says
Send a small amount first. Confirm it arrives. Then send the rest. This costs one network fee, and it is the only check that tests the whole path — address, network, memo, and the recipient's system — at the same time. A green result here is not a substitute for it.
The four things to confirm before pressing send, and what is recoverable when something has already gone wrong, are in withdrawal fees by network.