Jump to content

Connect SuperML | Leeroopedia MCP: Equip your AI agents with best practices, code verification, and debugging knowledge. Powered by Leeroo — building Organizational Superintelligence. Contact us at founders@leeroo.com.

Principle:Infiniflow Ragflow Agent Execution

From Leeroopedia
Knowledge Sources
Domains RAG, Agent_Systems, Workflow_Orchestration
Last Updated 2026-02-12 06:00 GMT

Overview

A graph-based workflow execution engine that traverses the agent DSL, invokes components in topological order, and streams results via SSE events.

Description

Agent Execution is the runtime engine that processes the agent's DSL graph. The Canvas.run async generator traverses the graph from the Begin node, invoking each component in order based on edge connections. It handles branching (Categorize/Switch), iteration (Loop/Iteration), user input pauses (UserFillUp), error handling with exception handlers, and streams SSE events (workflow_started, node_started, node_finished, message, message_end, user_inputs, workflow_finished).

Usage

Triggered via the POST /canvas/completion endpoint when a user sends a query to an agent.

Theoretical Basis

The execution model follows a directed acyclic graph (DAG) traversal with extensions:

  • Topological ordering: Components execute based on edge dependencies
  • Streaming: LLM components yield partial results for real-time display
  • Branching: Categorize/Switch nodes direct flow based on content classification
  • Iteration: Loop/Iteration nodes enable repeated execution patterns

Related Pages

Implemented By

Page Connections

Double-click a node to navigate. Hold to expand connections.
Principle
Implementation
Heuristic
Environment