Skip to main content

​
Tool Edges

Purpose: Connect Agent nodes to Tool nodes for AI tool calling Tool edges are special connections that make tools available to Agent nodes. When an Agent is connected to Tool nodes via tool edges, those tools become part of the Agent’s available function set. Tool Edges - Agent node connected to Tool node via TOOLS handle In this example:
  • The Agent Node has a TOOLS output handle
  • A tool edge (dashed line with 🔧 wrench icon) connects to the Tool Node
  • The tool becomes available to the Agent for function calling
  • The Agent can decide when and how to use this tool during its iteration loop
Other handles visible:
  • PARALLEL — For standard execution flow (to downstream nodes)
  • ERROR — For error handling
How It Works:
1

Connection

Connect Tool nodes to an Agent node’s TOOLS handle using tool edges
2

Discovery

The Agent automatically discovers all connected tools at the start of its loop
3

Decision

During execution, the LLM decides which tools to call based on the conversation
4

Execution

Tools execute and return results to the Agent
5

Iteration

The Agent can call tools multiple times before producing a final response
Visual Indicator:
  • Icon: 🔧 Wrench
  • Label: Tool name or icon
  • Connection: From Agent’s TOOLS handle to Tool node’s input
Example Use Case: An AI customer support agent with three tools:
  • Search Knowledge Base — Find help articles
  • Check Order Status — Query order database
  • Create Ticket — Generate support ticket
The Agent can use any combination of these tools to answer the user’s question.
Tool edges are not execution edges — they don’t trigger node execution directly. They only define which tools are available to the Agent. The Agent handles tool execution internally during its iteration loop.

​
What’s Next?

Agent Node

Learn how Agents use tools in their iteration loop

Tool Node

Explore available tool types and MCP servers

Parallel Edges

Standard execution flow between nodes