How agents work

YieldSeeker agents are autonomous onchain systems with a user-defined purpose. The agent type determines its goal and strategy; the surrounding account, wallet, authorization, and safety model are shared.

The shared lifecycle

1. Connect a wallet

Your connected wallet identifies you and remains the owner of your YieldSeeker participation.

2. Create an account and agent

A YieldSeeker account can contain up to two active agents on the Free plan or five on Pro. When you create an agent, you choose its type and the configuration supported by that type.

3. Configure the goal

For the current Vault Agent, you can set an investment profile and natural-language preferences. Future agent types may expose different settings.

4. Fund or activate the agent

The current Vault Agent must hold at least $10 worth of its base asset and be deployed before it can allocate assets.

5. Let the agent operate

An active agent evaluates its permitted opportunities and takes actions according to its type-specific strategy. It cannot bypass the platform’s shared execution restrictions.

6. Monitoring watches your rules

Your profile and preferences become rules. Where a rule has a measurable limit, the platform turns it into boundaries on each position the agent holds and checks them continuously against live data. For the current Vault Agent, that covers limits on a vault’s available liquidity, its total deposits, and the collateral it uses.

Each boundary is clear, approaching (close to your limit, or trending toward it), or breached. A breach wakes the agent: it becomes eligible for its next run without waiting for its usual schedule, so it can re-evaluate that position. Monitoring does not move funds or send notifications itself; you can see each boundary’s status on the agent’s overview.

7. Review, adjust, or exit

You can review the results and activity available for the agent type, change supported settings, withdraw where supported, or stop participating. When you change your preferences, the rules update and monitoring starts watching the new boundaries.

What changes between agent types?

The type determines:

the objective the agent is pursuing
the assets and networks it supports
the actions it can request
the limits monitoring can watch
the metrics that describe performance
the risks you need to understand

For example, the Vault Agent pursues risk-adjusted yield across approved vaults, while the coming-soon Borrow Agent will manage a loan against your crypto. Each type gets its own guide under Agent types.