Set budgets, define spending policies in plain English, and approve or reject AI agent purchases in real time — before the money leaves your account.
Agents call APIs, provision cloud resources, and subscribe to SaaS tools — often without explicit limits.
You find out about costs after the fact — when the invoice arrives. By then it's too late.
Prompt instructions like "don't spend more than $50" are suggestions, not guarantees. Agents can ignore them.
LetAgentPay sits between your AI agent and the payment — a policy checkpoint that enforces your rules.
Set budgets, category limits, schedules, and per-request caps — in plain English or JSON.
Before spending, your agent calls LetAgentPay with the amount, category, and description.
8 checks run instantly: budget, category, limits, schedule, spending history. Auto-approve or escalate.
Approved requests proceed. Flagged ones wait for your review. Rejected ones stop the agent.
We're not a payment processor. We're the policy layer that decides whether a payment should happen.
The agent asks LetAgentPay for approval before initiating any transaction. No approval — no payment.
Policies are checked instantly. Auto-approve for routine purchases, escalate unusual ones to you.
The agent confirms the result. Spending counters update. Full audit trail preserved.
Write rules like "max $100/day on cloud services, no weekend spending". AI converts them to enforceable JSON policies.
Set per-request, daily, weekly, monthly limits and an overall budget. All checked in real time against actual spending.
Restrict agents to specific spending categories — groceries, cloud, SaaS, travel — and set different limits per category.
Allow spending only during business hours, block weekends, or define custom time windows.
Routine purchases within policy get approved instantly. Unusual or high-value requests require your manual review.
See every request as it happens. Approve, reject, or review with full context — amount, category, merchant, policy check results.
Track daily, weekly, and monthly spending by agent. See budget utilization and remaining balances at a glance.
Every request, every policy check, every decision — logged and searchable. Know exactly what was spent, when, and why.
Authorize on-chain USDC payments on Base. Same policy engine for fiat and crypto — unified budget, one dashboard.
Three ways to connect your agent — pick what fits your stack.
pip install letagentpay
client = LetAgentPay(
token="agt_..."
)
client.purchase(
amount=49.99,
category="cloud"
)Works with Claude Desktop, Cursor, and any MCP-compatible AI tool. One config line — your agent gets spending tools automatically.
Simple HTTP endpoints with Bearer token auth. Works with any language or framework. Full OpenAPI spec available.
Give your users confidence that their agent won't overspend. Add spending controls as a feature, not a liability.
Let agents handle procurement, cloud provisioning, and SaaS subscriptions — with guardrails you define.
Trust your personal AI to book, shop, and subscribe — knowing every purchase gets checked against your rules first.
Free while in early access. Set up in under 5 minutes.