corellis
Corellis is an open-source multi-agent governance framework that enables scaling OpenClaw from a single AI assistant to a coordinated fleet of 20 or more agents. It provides core capabilities for goal decomposition, allowing high-level objectives to be broken into sub-goals that can be distributed across agents. Fleet-wide memory ensures that all agents share a common knowledge base, enabling collective recall and consistent context. Correction propagation mechanisms route feedback and corrections to the relevant agents, helping the fleet learn and adapt. Approval workflows introduce human oversight and controlled delegation, ensuring that agent actions can be reviewed before execution. Designed for complex AI agent systems, Corellis addresses the challenges of coordination, knowledge sharing, and error correction in multi-agent environments. By providing shared knowledge and collective memory, the framework reduces duplicated efforts and maintains alignment across distributed goals. The open-source nature of the project allows developers to inspect, modify, and extend the governance logic to suit specific use cases. Corellis is intended for teams and developers who need to manage large fleets of AI assistants with robust governance and oversight.
- Category
- AI Agents
- Pricing
- Open Source
Problem solved
Managing large fleets of AI agents is difficult due to coordination overhead, inconsistent knowledge, and lack of governance. Corellis solves this by providing a governance framework that decomposes goals, maintains fleet-wide memory, propagates corrections, and enforces approval workflows, enabling reliable operation of 20+ agent fleets.
Use cases
- Coordinating a fleet of AI assistants for complex task execution
- Distributing goals across multiple agents with shared context
- Propagating corrections to agents to improve collective performance
- Implementing approval workflows for agent actions
- Maintaining fleet-wide memory for knowledge retention
- Scaling from a single AI assistant to 20+ agent fleets
Audience
- Developers
- ML Engineers
- Data Scientists
- Researchers
Pros
- Enables coordination of large agent fleets (20+)
- Provides goal decomposition and fleet-wide memory
- Supports correction propagation for continuous improvement
- Includes approval workflows for human oversight
- Open source and free to use
Cons
- Requires OpenClaw platform (may limit adoption)
- Documentation or community support may be limited
- Complexity of managing multi-agent governance
- Early stage or limited production validation
Official website: https://github.com/CorellisOrg/corellis
Canonical LambdaBase page: https://www.lambdabase.com/tools/corellis