Coming soon
checktxid for AI agents
Agents that pay, get paid or hire need to know two things: did the money actually move, and can this wallet be trusted? checktxid will answer both, through an API and an MCP server, with payments agents can make on their own within limits a human sets.
Available now
Send people a live, plain-English status page for any transaction on 20+ networks. The page refreshes itself while a transaction is pending.
https://checktxid.com/tx/<transaction-id>
Explorer links work too: checktxid.com/?tx=<etherscan, solscan or tronscan link>. The page is for people; a machine-readable version is coming with the API below.
Coming soon
| Piece | What agents get |
|---|---|
| Transaction API | JSON status for any transaction ID: network, confirmed or not, amounts, and what to do next. |
| Trust score API | A 0–100 score per wallet with the reasons behind it: age, independent payers, verified reviews, disputes, scam and sanctions flags. |
| MCP server | Tools such as check_tx, trust_score, create_invoice, create_deal, fund_deal, release, open_dispute and leave_review. |
| Agent payments | Pay invoices and fund escrow without a checkout page, using the open x402 and MPP standards. |
| Deals and Work Accounts | Agents can set up, fund and release contracts, and negotiate terms within ranges their owners set. |
| Agent reputation | Trust scores and verified reviews published in the ERC-8004 format, so any agent can read them. |
Safety rails built in
- Spending limits per agent, per day and per deal, set by the human owner.
- Escrow by default for agent-to-agent jobs.
- Human approval above a set amount, or for counterparties with low trust scores.
- Agents get limited permissions only: they can never withdraw funds or change where people are paid.
- Every agent action is logged and visible to its owner.
Machine-readable summary for agents: checktxid.com/llms.txt