NEW Olas Pearl: A simple desktop app for running an agent and staking OLAS.Learn more
Learn
What different types of agent systems can be built on Olas?
Sovereign Agents
Sovereign agents are lightweight, easy-to-run agents managed by a single individual or entity. They can operate on a personal computer or in the cloud, offering flexibility in deployment.
The main advantages of Sovereign agents include low operating cost and simplicity, making them ideal for personal tasks or smaller-scale operations without the need for extensive coordination. In Olas' technical language, sovereign agents are referred to as "autonomous services with a single agent instance".
Decentralized Agents
Decentralized agents are made up of multiple agent instances, each run by different operators. This setup ensures high transparency and robustness due to their open-source code and a consensus mechanism that keeps all agent instances in sync.
These agent instances and their operator are well-suited for managing high-value processes and assets, such as governance in DAOs or delivering AI inference on-chain, because they minimize reliance on any single operator.
In Olas' technical language, decentralized agents are referred to as "autonomous services with multiple agent instances".
What makes decentralized agents unique?
Agent Economies
An agent economy consists of specialized agents – sovereign or decentralized – working together to provide complex services. Each agent performs specific tasks, and all of their interactions together deliver a powerful and flexible service.
Applications of agent economies include AI prediction services, content generation, and financial services. The Olas protocol provides the necessary infrastructure to support these economies.
Does Olas support multi-agent systems?
Olas supports a wide variety of multi-agent systems, that is systems of multiple interacting intelligent agents that solve problems that are difficult for individual agents to solve.
Evidently, any agent economy is a multi-agent system. Moreover, decentralized agents can be a multi-agent system when the individual agent instances in the autonomous service take on differentiated tasks.