The Problem
AI agents are powerful, but giving them direct access to your crypto wallets is dangerous.
- ✗Private keys exposed to agent code
- ✗No spending limits or controls
- ✗Single point of failure
- ✗No audit trail of agent actions
# Dangerous pattern
wallet = Wallet(private_key)
agent.wallet = wallet # Full access!The Solution
ZeroQuant vaults act as a secure intermediary between your AI agents and DeFi protocols.
Built for Developers
First-class support for Python and TypeScript. Native LangChain tools included.
from zeroquant import ZeroQuantClient
from zeroquant.langchain import (
CreateVaultTool,
ExecuteSwapTool
)
from langchain.agents import AgentExecutor
# Initialize client
client = ZeroQuantClient(
rpc_url="https://arb-mainnet.g.alchemy.com/v2/...",
signer=your_signer
)
# Create LangChain tools
tools = [
CreateVaultTool(client),
ExecuteSwapTool(client),
]
# Build your agent
agent = AgentExecutor(
agent=your_llm_agent,
tools=tools
)import { ZeroQuantClient } from '@zeroquant/sdk';
import {
createVaultTool,
executeSwapTool
} from '@zeroquant/langchain';
// Initialize client
const client = new ZeroQuantClient({
rpcUrl: 'https://arb-mainnet.g.alchemy.com/v2/...',
signer: yourSigner
});
// Create LangChain tools
const tools = [
createVaultTool(client),
executeSwapTool(client),
];
// Build your agent
const agent = new AgentExecutor({
agent: yourLlmAgent,
tools
});Everything You Need
Battle-Tested Security
Smart contracts built with OpenZeppelin. Audited and battle-tested on mainnet.
Lightning Fast
Optimized for L2s. Batch operations and gas-efficient execution.
Multi-Chain
Deploy on Arbitrum, Optimism, Base, Polygon, and any EVM chain.
Developer First
TypeScript + Python SDKs with full type safety and comprehensive docs.
Agent Native
Built-in LangChain tools. Works with any agent framework.
Full Observability
Transaction logs, spending analytics, and real-time monitoring.