sipi.bot vs Lago billing
Lago is an open-source usage-based billing engine. It meters and prices usage events. It does not evaluate whether the event should have occurred.
What Lago does
Lago (AGPL) ingests events, applies pricing plans, and computes invoices. For agent spend, Lago's role is post-hoc: it records usage that already happened. The pre-spend decision — whether the agent should have been allowed to make that call — is a separate layer.
Where they differ
| Dimension | Lago | sipi.bot |
|---|---|---|
| Stage | Post-spend (metering) | Pre-spend (policy gate) |
| Decision | Records & prices | APPROVE/BLOCK/FLAG in <5ms |
| Loop protection | Prices the loop | Blocks at first violation |
| Audit trail | Billing records | Tamper-evident per-tx log |
| Human escalation | N/A | FLAGGED → approval queue |
Frequently asked questions
Do I need both Lago and sipi.bot?
Yes. Billing platforms meter what was spent; sipi.bot gates what can be spent before it occurs. For autonomous agents, you need the billing engine for pricing and the firewall for policy enforcement.
Can sipi.bot integrate with Lago?
The two layers operate independently. Your agent calls evaluate_spend before the transaction; if approved, the billing platform meters and prices the usage normally.
What's the failure mode without a firewall?
A retry loop generates valid usage events continuously. Without a pre-spend check, every event is metered, priced, and invoiced at full accuracy. The bill is correct — just catastrophically large.