Ruflo
An open-source AI agent framework that coordinates Claude Code and Codex, with persistent memory, Ollama support and collaboration across machines.
Ruflo is an AI agent framework for developers who want Claude Code and Codex to work as coordinated teams. It adds specialist roles and shared context for projects that involve coding, testing, architecture and documentation.
- Agent teams: Divide work among specialist agents and coordinate their tasks and decisions, with background workers for audits, optimization and finding test gaps.
- Persistent memory: Keep agent context across sessions and retrieve it through AgentDB vector search, hybrid search and knowledge graphs.
- Learning from feedback: Use task histories and feedback to inform later work, with results dependent on the configured memory and learning components.
- Cross-machine collaboration: Let agents on separate machines exchange work with identity verification, trust controls, outbound personal-data filtering and audit records.
- Review tools: Plugins scan for vulnerabilities and prompt injection, while MetaHarness checks agent configurations and tracks changes to catch regressions.
Ruflo is open source under the MIT license. It supports Ollama for local LLM use alongside Claude, GPT, Gemini and Cohere. Choosing a cloud model provider sends model requests to that provider; cross-machine collaboration also exchanges data beyond the local instance.