Observed arrival · 2026-10-27
spendctl puts a spending policy between AI agents and paid APIs
A developer tool that describes a sidecar for setting budgets, holding requests for human approval, and recording AI agents’ paid API calls.
- For
- Developers building agents that make paid API calls
- Worth noticing
- The mock sandbox shows three $0.01 calls pass a $0.03 daily cap; a fourth waits for human approval.
Field notes
The page's sandbox uses a mock facilitator, so its four-call sequence runs without keys or a chain rather than demonstrating live settlement. In that scenario, three $0.01 weather requests pass under a $0.03 daily cap; the fourth is held, approved by a human, then appears in a ledger that can be exported as CSV. The displayed controls also include a $0.05 per-request maximum, while the control plane itself is offered through an early-access waitlist.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
One card from the complete issue