sardis-langchain

LangChain integration for Sardis -- the Payment OS for the Agent Economy.

Gives your LangChain agents the ability to make real, policy-enforced stablecoin payments through Sardis MPC wallets.

Installation

pip install sardis-langchain

Quick Start

from sardis import SardisClient
from sardis_langchain import SardisToolkit

# 1. Initialize Sardis
client = SardisClient(api_key="sk_...")
wallet = client.wallets.create(
    name="procurement-agent",
    chain="base",
    policy="Max $100 per transaction, daily limit $500",
)

# 2. Create LangChain tools
toolkit = SardisToolkit(client=client, wallet_id=wallet.id)
tools = toolkit.get_tools()

# 3. Attach to any LangChain agent
from langchain.agents import AgentExecutor, create_react_agent
from langchain_openai import ChatOpenAI

llm = ChatOpenAI(model="gpt-4o", temperature=0)
agent = create_react_agent(llm, tools, your_prompt)
executor = AgentExecutor(agent=agent, tools=tools, verbose=True)

result = executor.invoke({
    "input": "Pay $30 to Anthropic for API credits and $15 to GitHub for Copilot"
})

Available Tools

Toolkit Options

Full toolkit (all 5 tools)

toolkit = SardisToolkit(client=client, wallet_id=wallet.id)
tools = toolkit.get_tools()

Payment-only tools (3 tools)

tools = toolkit.get_payment_tools()  # pay + balance + policy check

Individual tools

from sardis_langchain import SardisPayTool, SardisCheckBalanceTool

pay_tool = SardisPayTool(client=client, wallet_id=wallet.id)
balance_tool = SardisCheckBalanceTool(client=client, wallet_id=wallet.id)

Callback Handler

Track Sardis operations with the built-in callback handler:

from sardis_langchain import SardisCallbackHandler

handler = SardisCallbackHandler()
executor = AgentExecutor(agent=agent, tools=tools, callbacks=[handler])

# After execution, inspect the audit trail
for record in handler.get_payment_records():
    print(record)

How Policy Enforcement Works

Every payment goes through Sardis's policy engine before execution:

Agent calls sardis_pay(to="openai.com", amount="25.00")
    |
    v
Policy Engine checks:
    1. Per-transaction limit  -> Is $25 under the cap?
    2. Balance check          -> Enough funds?
    3. Total spending limit   -> Under lifetime cap?
    4. Wallet active          -> Is the wallet enabled?
    |
    v
PASS -> Execute payment, return tx_hash
FAIL -> Return rejection reason (no funds move)

Simulation Mode

By default, SardisClient runs in simulation mode -- all operations execute locally without network calls. This is ideal for prototyping and testing agent payment flows.

# No API key needed for simulation
client = SardisClient()
wallet = client.wallets.create(name="test-agent", chain="base", policy="Max $50/tx")

toolkit = SardisToolkit(client=client, wallet_id=wallet.id)
# All tools work locally -- no real money moves

Links

• Sardis Documentation
• GitHub Repository
• LangChain Documentation