Use Cases
The coordination layer for autonomous agents. When your AI systems need to elect a leader, vote on decisions, or prevent conflicts — Arbiter provides the infrastructure for machines to reach agreement.
When Agents Need a Single Voice
The Scenario
You're running a fleet of trading agents. They each have their own strategies, but they share a single portfolio. Who decides the overall position?
Without Arbiter: You designate one agent as "master" — creating a single point of failure. If it crashes, goes rogue, or gets exploited, everything stops.
With Arbiter: The agents elect a leader for a fixed term. If the leader stops responding, a new election happens automatically. The leader's authority is time-bounded and revocable. Every trade the leader makes includes proof of their mandate.
When Consensus Matters
The Scenario
Your security monitoring agents have detected unusual behavior from one of their peers. Is it compromised? Should they isolate it?
Without Arbiter: Each agent makes its own decision. Some isolate the suspect; others don't. The result is inconsistent and exploitable.
With Arbiter: The agents vote. The proposal: "Isolate Agent-7 for suspicious behavior." If the vote passes, the decision is binding and recorded. If Agent-7 disputes it later, the evidence is on-chain.
When Resources Can't Be Shared
The Scenario
Multiple agents need to update a shared database. Concurrent writes would corrupt the data.
Without Arbiter: You implement locking at the database level. But what if an agent acquires a lock, then crashes? The lock is held forever.
With Arbiter: Locks expire automatically. Each lock includes a "fencing token" — a number that only increases. The database rejects writes from agents holding stale tokens. Even if a crashed agent "wakes up," it can't corrupt data.