Auto executes complex, multi-step workflows from a single request — gathering data, making recommendations, and executing trades in sequence. You do not need to break tasks into individual commands because Auto carries context across steps and conversation turns.
What Is a Multi-Step Operation?
A multi-step operation is a request where Auto needs to:
Gather data first (balances, prices, research)
Make a decision or recommendation based on that data
Execute one or more actions (swaps, bridges, deposits)
Carry context across multiple conversation turns
This differs from simple commands because Auto chains dependent steps together automatically.
Can Auto Handle Multiple Actions in One Message?
Auto completes several actions from a single prompt when the steps have a clear sequence.
"Check my ETH on Base and Arbitrum, swap 30% of my Base ETH to USDC, then bridge 200 USDC from Base to Ethereum."
Auto handles the steps in order: checks balances first, then swaps, then bridges.
How Does Research-Then-Trade Work?
A common multi-step pattern is researching first, then acting on the result across multiple turns.
Example: Trending Memecoin to Swap
"What memecoins are trending on Base right now? Show top 3 with quick risk notes."
Auto returns the top choices with risk analysis.
"Buy $300 of the strongest one using USDC on Base."
Auto resolves "the strongest one" from the previous response and executes the swap.
Example: Research, Bridge, Then Swap
"I have 2 ETH on Ethereum. What's trending on Base and what should I rotate into?"
After Auto gives a suggestion:
"Bridge 0.6 ETH to Base."
Then:
"Once it lands, swap 0.4 ETH into that token."
Auto remembers the recommended token, waits for the bridge to settle, and completes the trade.
Example: Compare Assets, Then Buy the Winner
"Compare the last 30 days for ARB vs OP and tell me which has better momentum."
Auto compares price action and suggests the stronger setup.
"Okay, buy $500 of your pick on Arbitrum."
The follow-up does not need to repeat the token name — Auto resolves it from the earlier analysis.
Example: Portfolio Review to Rebalance
"Review my wallet across Ethereum, Base, and Arbitrum and suggest a rebalance to 40% ETH, 40% stables, 20% high beta."
Auto inspects the portfolio and proposes exact trades.
"Do the rebalance on Base first."
That follow-up moves Auto from analysis into execution.
How Does Auto Carry Context Across Turns?
Auto resolves references from earlier messages when the conversation is specific and recent. You can use shorthand follow-ups:
"Buy the strongest one"
"Do that on Base"
"Pause that task"
"Use the highest-liquidity one"
Auto resolves each reference from the earlier context as long as the reference is unambiguous.
What Are the Best Use Cases for Multi-Step Operations?
How Do I Get the Best Results from Multi-Step Requests?
Be specific about chain, amount, or venue when it matters
Ask for research first if you want a recommendation before acting
Use natural follow-ups instead of repeating the entire request
When Auto presents multiple options, choose one clearly in the next turn
Example Prompts to Try
"Compare BTC, ETH, and SOL momentum, then tell me which looks strongest."
"Check my wallet, find the best chain for a USDC swap, then execute it."
"Show me active Polymarket markets on the Fed's next rate decision, then buy $150 YES on the highest-liquidity one."
"Review my portfolio and suggest a rebalance. If the Base side needs work, do that first."
