fizzl

Safety checks for Python AI agents that pay: LangChain, LangGraph, the OpenAI Agents SDK, CrewAI or your own loop.

The checks only check. They never sign, pay or move anything themselves. A failed check comes back as {"error": ..., "message": ...} instead of raising, so the agent can tell its user.

Install

pip install "fizzl[x402]"              # with the x402 client, to pay per check
pip install "fizzl[langchain,x402]"    # plus LangChain tools
pip install "fizzl[openai-agents]"     # OpenAI Agents SDK tools (Python 3.10+)

Use

import requests
from eth_account import Account
from x402 import x402ClientSync
from x402.http.clients import wrapRequestsWithPayment
from x402.mechanisms.evm.exact import ExactEvmScheme
from fizzl import Fizzl

payer = x402ClientSync().register("eip155:8453", ExactEvmScheme(Account.from_key(AGENT_KEY)))  # USDC on Base
session = wrapRequestsWithPayment(requests.Session(), payer)

fizzl = Fizzl(session=session)
fizzl.check_endpoint_before_paying("https://api.example.com/funding", max_usd=0.05)
# {'verdict': 'go', 'summary': 'OK to pay: $0.02 on Base.', 'options': [...], 'reasons': [], ...}
fizzl.check_token("base", "0x...")
fizzl.check_before_signing(type="approval", chainId=8453, token="0x...", spender="0x...", amount="115792089237316195423570985008687907853269984665640564039457584007913129639935")

Try it without a wallet: Fizzl() with no paying session and no credit keys falls back to the free quick checks for check_token, check_before_signing and check_endpoint_before_paying: the verdict only, a few per hour, marked "free": True with a note on what the full check costs. check_wallet_approvals has no free version. Fizzl(free=False) returns payment_required instead.

Prepaid credits instead of a payment per check: buy a pack once (presign-guard: 100 checks for $0.80; x402 Doctor: 1000 preflights for $0.80), then use a plain session:

fizzl = Fizzl(credit_keys={"presign": PRESIGN_CREDIT_KEY, "doctor": DOCTOR_CREDIT_KEY})

LangChain / LangGraph

from langgraph.prebuilt import create_react_agent
from fizzl.langchain import fizzl_tools

tools = fizzl_tools(session=session)          # or credit_keys={...}; only=[...] for some of them
agent = create_react_agent(model, tools + your_tools,
    prompt="Before you pay any API, call check_endpoint_before_paying. Before you sign anything, call check_before_signing. Never continue on red or no_go.")

The tools are plain StructuredTools with typed arguments, so CrewAI and other frameworks that take LangChain tools can use them too.

OpenAI Agents SDK

from agents import Agent, Runner
from fizzl.openai_agents import fizzl_tools

agent = Agent(name="buyer", tools=fizzl_tools(session=session),   # or credit_keys={...}, or nothing for the free checks
    instructions="Before you pay any API, call check_endpoint_before_paying. Before you sign anything, call check_before_signing. Never continue on red or no_go.")
print(Runner.run_sync(agent, "Is https://api.example.com/funding safe to pay, at most $0.05?").final_output)

The checks run in a worker thread, so a blocking requests session doesn't stall the agent's event loop. Bad arguments from the model come back as {"error": "bad_input", ...} without calling (or paying for) the check.

Want the wallet to enforce it, not just inform the model?

Tools inform the model; a model can still ignore them. To make red and over-budget payments impossible, guard the wallet itself: presign-guard-wallet (Node) checks every signature with presign-guard and your spending limits, with Telegram approval above them.

License

MIT