Many AI agents, shared company knowledge: Sakana introduces Aquarium
3-line summary
- Sakana AI announced Aquarium, a platform for agents sharing company knowledge and rules.
- It structures data relationships and supports human approvals and execution records.
- The official pages invite deployment inquiries without listing prices or a general-access date.
Sakana AI announced Aquarium on October 23 as a platform to build, run and operate workplace agents. The company says prepared data connections and knowledge structures can be reused by other agents, reducing repeated groundwork for each workflow.
Its foundation is an ontology: shared definitions of knowledge, rules, data and their relationships. The product page says connections and synchronization with existing data systems are designed for each deployment. Workflows can incorporate human approvals alongside information retrieval and execution.
The announcement describes access permissions, pre-execution approval settings and records linking decision evidence to results. The public product page invites deployment inquiries; it provides no price list, general-access date or quantitative performance comparison. This article does not report a hands-on product test.
Why it matters
AI Ping sees potential to reduce repeated preparation as teams add agents for different tasks. Actual benefits depend on how data connections and approval steps are designed.
Who it's for
- Developers
- Before a deployment discussion, map data sources, synchronization requirements and approval points.
- Office workers
- Choose one recurring task and separate agent actions from human review.
