Integrate sipi.bot with LangChain
LangChain agents call tools; tools spend money. Adding the guard is a few lines — here's the exact walkthrough.
Step 1 — Get a key
Create an agent key from the dashboard (or use the evaluate endpoint anonymously on self-host).
Step 2 — Add the guard tool
Register sipi.bot as a tool the agent can call before any purchase.
Step 3 — Call it before spending
In the tool that spends, call the guard with amount, merchant, and category. Branch on APPROVED / BLOCKED / FLAGGED.
Step 4 — Start with three rules
A per-transaction cap, a merchant allowlist, and a velocity limit.
Guard call
import requests
from langchain.agents import tool
@tool
def guard_spend(amount: float, merchant: str, category: str) -> str:
r = requests.post(
"https://sipi.bot/v1/transactions/evaluate",
json={"amount": amount, "merchant": merchant, "category": category},
headers={"Authorization": "Bearer YOUR_KEY"},
)
return r.json()["decision"] # APPROVED | BLOCKED | FLAGGED
Deterministic, ~5 ms, no model call.
FAQ
Does this slow the agent?
No — about 5 ms per check, only on spend actions.
Is there a LangChain wrapper?
A client wrapper takes a few lines — or just use the HTTP call above.
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