AgentPays issues a single-use virtual card scoped to each approved purchase, so a prompt-injected or compromised agent leaks one disposable card, not your funding source.
All purchase requests across your agents, most recent first.
| Date | Agent | Merchant | Purchase | Fee(1.5%, $0.50 min) | Total | Justification | Status | Card Status | Card | Fee Card | Promo | Decided by | Decision reason |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| May 13, 2:32 PM | gpt-researcher | OpenAI API | $42.10 USD | $0.63 | $42.73 | Refill API credits for nightly research run | Approved | Active | ****4127 | ****8801 | — | Auto | Under auto-approve threshold; within daily limit |
| May 13, 11:08 AM | code-agent | GitHub Marketplace | $19.00 USD | $0.50 | $19.50 | Copilot seat for code-review subagent | Approved | Active | ****9034 | ****8801 | — | Auto | Merchant on allowlist; under monthly cap |
| May 12, 10:14 PM | eval-runner | Anthropic | $248.00 USD | $3.72 | $251.72 | Top up Claude API balance for batch eval | Approved | Active | ****5512 | ****8801 | LAUNCH25 | Kevin Fee | Over auto-approve threshold; approved via email |
| May 13, 3:47 PM | web-scraper | ScrapingBee | $8.40 USD | $0.50 | $8.90 | Extra credits for competitor pricing crawl | Pending review | — | — | — | — | — | — |
| May 12, 9:21 AM | data-buyer | unknown-merchant.io | $612.00 USD | — | $612.00 | Premium dataset license | Denied | — | — | — | — | Auto | Merchant not on allowlist; exceeds per-txn limit |
Showing 5 most recent requests
Disposable cards, server-side rules, and human approval when it matters.
The agent never touches your credit card, bank account, or Privacy.com key. Each approved purchase gets a single-use virtual card delivered through a time-limited view link, so a compromised agent leaks one disposable card and nothing else.
A server-side rules engine checks every request against budgets, merchant allowlists, and velocity limits before a card is ever issued. Anything above your auto-approve threshold pauses for an email approval from you, so policy and human judgment both gate the spend.
Claude Desktop, Claude.ai, and ChatGPT connect over OAuth with no API key to manage. Claude Code, Codex, and custom agent frameworks authenticate by passing the per-agent API key as a bearer token in the MCP config. Same server, same tools, either way.
Create an agent, set your spending rules, and ship a working integration in minutes. Your real card stays with you.
Issue your first virtual card