[ SHORT ANSWER ]
AgentKit is a good framework and you do not have to leave it — the gap is that its quickstart path puts a wallet the agent can use directly into the agent's own process. 1Claw sits underneath as the signer: the agent keeps its AgentKit tools, but the key lives server-side and every transaction passes destination allowlists, value caps, and optional human approval first.
The risk with any agent framework is that the convenience path and the safe path differ. AgentKit's convenience path hands the agent signing authority. Because 1Claw's guardrails evaluate the decoded inner call, they still apply when the transaction is wrapped in a Safe or an ERC-4337 user operation — which is where naive policy layers tend to be bypassed.
[ COMPARISON ]
1Claw vs Coinbase AgentKit (raw keys)
Capability by capability. A dash means partial or qualified support — read the note.
[ WHEN TO USE ]
Which one is right for you
Most of these are not either/or. Where the competitor is the better answer, we say so.
Use Coinbase AgentKit (raw keys) when
- A hackathon or prototype on testnet where speed is the only goal.
- You want Base-native actions and no additional dependency.
- The agent's wallet holds an amount you are willing to lose entirely.
Use 1Claw when
- The agent's wallet holds real value on mainnet.
- You need caps and allowlists that survive Safe or 4337 wrapping.
- You want one audit trail across the agent's transactions and its API credentials.
[ MIGRATION ]
Moving over in three steps
Most teams keep Coinbase AgentKit (raw keys) for what it is good at and add 1Claw for the agent layer.
- 1Provision a 1Claw signing key for the agent and fund it with an operational balance only.
- 2Configure allowlists and per-chain daily budgets before pointing the agent at mainnet.
- 3Swap AgentKit's local signer for the Intents API so signing happens server-side under policy.
[ FAQ ]
Common questions
Do I have to stop using AgentKit?
No. The change is which component holds the key and enforces policy, not which framework you build the agent in.
Does 1Claw work on Base?
Yes — Base is an EVM chain and is covered by the same signing, guardrail, and simulation path as Ethereum.
Can guardrails be bypassed by wrapping the call?
No. Guardrails run against the decoded inner call before Safe or ERC-4337 wrapping is applied, specifically so that wrapping is not an escape hatch.
[ RELATED ]
Other comparisons
[ FREE TIER ] 3 vaults · 50 secrets · 2 agents · 100 signatures/mo
Give agents access, not copies
Store the credential once, scope it per agent, and let the agent act without ever holding it.