AI Agents

Tool Calling

Tool calling lets AI agents interact with the real world — running queries, calling APIs, reading files, and more — without leaving the workflow.

How It Works

1. You define Tools by connecting nodes to the agent's tool inputs
2. The agent decides *when* and *how* to call each tool based on the task
3. riogentix executes the tool and returns the result to the agent
4. The agent uses the result to continue reasoning

Defining a Tool

Any action node can become a tool. Give it a:

• Tool Name — short, snake_case identifier (e.g. get_customer_orders)
• Description — tells the agent when to use it (be specific!)
• Input Schema — fields the agent must provide

Example: Customer Lookup Tool

Tool Name: lookup_customer

Description: Look up a customer by email address. Returns customer ID, name, plan, and account status.

Underlying Node: HTTP Request to your CRM API

Input Schema:

{
  "email": { "type": "string", "description": "Customer email address" }
}

Built-in Tools

riogentix provides pre-built tools you can attach directly:

• Web Search — search the internet
• Calculator — evaluate math expressions
• Date/Time — get current date/time with timezone support
• Code Interpreter — run Python snippets in a sandbox
• File Reader — read files from cloud storage

Tool Execution Trace

In the Execution Inspector, expand any agent step to see:

Step 3: Tool Call
  Tool: lookup_customer
  Input: { "email": "alice@example.com" }
  Output: { "id": "cus_123", "plan": "Enterprise", "status": "active" }
  Duration: 142ms