Copilot for Portfolio & Transactions

Portfolio & transaction copilot. Ask questions & execute via natural language.

Capabilities

Portfolio Queries
Portfolio Queries

Ask questions about your portfolio, balances, & holdings across all chains.

"What's my net worth?"
"Show portfolio breakdown"
Transaction Execution
Transaction Execution

Execute trades, transfers, & DeFi operations through natural language.

"Trade 100 USDC for ETH"
"Send 0.5 ETH to vitalik.eth"
Market Intelligence
Market Intelligence

Get real-time market data, price comparisons, & yield information.

"ETH price?"
"USDC lending APY rates"

Function Calling

The AI assistant uses function calling to interact directly with blockchain protocols.

functions.ts
// Portfolio Functions
get_wallet_net_worth(address)  // Multi-chain balances & total value
getUserProfile()               // Get user preferences

// Trade Functions
getTradeRoute()                // Get trade routes
executeTrade()                 // Execute transactions

// User Functions
saveUserPreference()           // Save preferences

Response Time & Data

Quick vs full response: Simple questions get fast answers. Portfolio & multi-chain queries use tools & can take 10–30+ seconds; the assistant does not stream. Optional memory helps the assistant remember context across the conversation but can add a bit of delay.

Why might my balance look different? Portfolio data can come from a single chain or from multi-chain totals (via get_wallet_net_worth). Numbers can differ depending on which source the assistant used & which chains are included.

If you see a generic error: Retry the message once. If it persists, check your connection & try rephrasing; if the assistant times out, ask a shorter question or try again later.

Safety Features

Confirmation Gates
Confirmation Gates

All high-risk actions require explicit user confirmation.

Risk Warnings
Risk Warnings

Alerts for high slippage or low liquidity operations.

No Assumptions
No Assumptions

The AI never invents numbers - only reports actual data.

Was this helpful?