Spend Control for Manus
Manus runs long-horizon autonomous tasks — browsing, executing, and spending across many steps. The guard bounds the run.
Why Manus runs overspend
Long-horizon autonomy means hundreds of tool calls per task.
Each step can trigger paid APIs or purchases.
No dollar-level budget inside the agent.
How it works
Call the guard before any spend-capable step: amount, merchant, category → APPROVED, BLOCKED, or FLAGGED.
Rules that fit
Per-task ceiling.
Velocity limit on step loops.
Merchant allowlist for paid tools.
Guard call
const res = await fetch("https://sipi.bot/v1/transactions/evaluate", {
method: "POST", headers: { "Content-Type": "application/json" },
body: JSON.stringify({ amount: 40, merchant: "api.vendor.com", category: "api" }),
});
const { decision } = await res.json();Long-horizon tasks, enforced budgets.
FAQ
Does this stop Manus autonomy?
No — it gates spend steps; the task continues within budget.
Can I cap per task?
Yes — per-agent rules per task.
Related
Stop the next $12,400 night.
One API call (or MCP tool) in front of every agent transaction — APPROVED, BLOCKED, or FLAGGED, deterministic, ~5 ms, fully logged.
See plans — from $99/mo Try a live check