Agent-to-agent payments

What it means here

An agent pays another agent the same way it pays anyone: with aidr pay to that agent's address. The payment comes out of the paying agent's Aider wallet, inside its limits, and lands as USDC at the other agent's address.

aidr pay 0x3C44…93BC 0.08 "Summary of the Q3 report (job 17)"

If agent B also has an Aider wallet, its owner can deposit what it earns into that wallet later. There is no direct wallet-to-wallet transfer inside the contract.

Both sides keep their own rules

• Agent A's payment must pass A's wallet rules: its per-payment and daily limits, recipient rules and pause switch. If B's address isn't allowed, A gets RecipientNotAllowed or RecipientBlocked.
• Agent B's own spending is limited by B's wallet, if it has one. Receiving money doesn't raise B's limits.

Memos that both sides can match

Put something both agents know in the memo, such as a job id: Summary of the Q3 report (job 17). The memo is stored in the AgentPaid event, so B, or B's owner, can find the payment on the explorer.

Checking that you were paid

The receiving agent can look up the payment by transaction hash on the explorer, or watch the AgentPaid events for its address; see watching activity. aidr history lists payments an agent made, not ones it received.

Don't pay the contract

Never pay the Aider contract's address thinking it is "agent B's wallet". Version 2 rejects that payment with InvalidRecipient and the CLI refuses it, because USDC sent to the contract that way would be stuck.

Next

See how to follow every payment in watching activity.