ruflo

Claude Code: With vs Without Ruflo

/docs/getting-into-the-flow/claude-code-with-vs-without-ruflo

CapabilityClaude Code AloneClaude Code + Ruflo
Agent CollaborationAgents work in isolation, no shared contextAgents collaborate via swarms with shared memory and consensus
CoordinationManual orchestration between tasksQueen-led hierarchy with 3 consensus algorithms (Raft, Byzantine, Gossip)
Hive Mindโ›” Not available๐Ÿ Queen-led swarms with collective intelligence, 3 queen types, 8 worker types
Consensusโ›” No multi-agent decisionsByzantine fault-tolerant voting (f < n/3), weighted, majority
MemorySession-only, no persistenceHNSW vector memory with sub-ms retrieval + knowledge graph
Vector Databaseโ›” No native support๐Ÿ˜ RuVector PostgreSQL with 77+ SQL functions, ~61ยตs search, 16,400 QPS
Knowledge Graphโ›” Flat insight listsPageRank + community detection identifies influential insights (ADR-049)
Collective Memoryโ›” No shared knowledgeShared knowledge base with LRU cache, SQLite persistence, 8 memory types
LearningStatic behavior, no adaptationSONA self-learning with sub-millisecond pattern matching, LearningBridge for insights
Agent ScopingSingle project scope3-scope agent memory (project/local/user) with cross-agent transfer
Task RoutingYou decide which agent to useIntelligent routing based on learned patterns (89% accuracy)
Complex TasksManual breakdown requiredAutomatic decomposition across 5 domains (Security, Core, Integration, Support)
Background WorkersNothing runs automatically12 context-triggered workers auto-dispatch on file changes, patterns, sessions
LLM ProviderAnthropic only5 providers (Anthropic, OpenAI, Google, Cohere, Ollama) with automatic failover and cost-based routing (cost-optimized routing)
SecurityStandard protectionsCVE-hardened with bcrypt, input validation, path traversal prevention
PerformanceBaselineFaster tasks via parallel swarm spawning and intelligent routing