Features

Everything you need to build and deploy AI agent workflows.

Agent Management

  • Create and configure AI agents with any LLM provider
  • System prompt layers with YAML-based composition
  • MCP tool integration with sandboxed sidecars
  • Agent-scoped memory with fact extraction

Graph Workflows

  • Visual node-based editor for multi-agent workflows
  • Supervisor routing with conditional branching
  • Parallel execution and loop support
  • Step-by-step execution tracing via SSE

Memory & RAG

  • Automatic fact extraction from conversations
  • Vector storage with Qdrant (hybrid dense + sparse)
  • Document upload, chunking, and embedding
  • Memory consolidation with importance decay

Platform Operations

  • Centralized agent and graph management
  • Multi-engine sync with automatic config polling
  • Engine registration and sync polling
  • Ops console for monitoring and diagnostics