Custom Tool Block

What it does: Create your own specialized tools that AI agents can use, extending their capabilities with custom functionality tailored to your needs.

When to Use This

Use Custom Tools when you need:

• ✅ Specialized functionality not available in built-in tools
• ✅ Custom business logic for AI agents
• ✅ Data transformations during agent workflows
• ✅ Integration with internal systems
• ✅ Reusable capabilities across multiple agents

Example: Create a "Calculate Quote" tool that takes product specs and returns a price estimate based on your custom pricing logic.

Key Features

• Flexible Design: Define exactly what inputs the tool accepts
• Custom Logic: Implement any functionality you need
• AI Integration: AI agents automatically know when to use your tool
• Reusable: Once created, use the tool in any workflow
• Self-Describing: Tool description helps AI understand when to use it

How Custom Tools Work

1. You define the tool: Describe what it does and what inputs it needs
2. You implement the logic: Define how the tool processes inputs
3. AI sees your tool: Agent understands the tool's purpose from your description
4. AI uses it naturally: When relevant, the agent calls your tool with appropriate inputs
5. Results flow back: Tool output is used in the agent's response

Setup Guide

Step 1: Define Your Tool

Give your tool a clear identity:

• Name: Short, descriptive identifier (e.g., "calculate_shipping")
• Description: Explain what the tool does and when to use it
• Example: "Calculates shipping cost based on package weight and destination"

Step 2: Define Input Parameters

Specify what information your tool needs:

Step 3: Implement the Logic

Define what your tool actually does when called:

• Process the input parameters
• Perform calculations or lookups
• Return the result

Step 4: Connect to AI Agent

Add your custom tool to an AI agent's toolkit. The agent will see your tool's name and description and use it when appropriate.

Common Use Cases

Business Calculations

Create tools for pricing calculations, discount logic, tax computation, or any formula-based operations.

Data Validation

Build tools that check if data meets your business rules - valid email formats, acceptable ranges, or required formats.

Formatting and Transformation

Convert data between formats, standardize inputs, or prepare data for other systems.

Lookup Operations

Create tools that fetch information from internal sources based on provided identifiers.

Custom Integrations

Connect to internal APIs or services that don't have built-in tool support.

Creating Effective Tools

Writing Good Descriptions

Your tool description tells the AI when to use it. Be clear and specific:

Good description: "Calculates the total shipping cost for a package. Use when a customer asks about shipping prices. Takes package weight in pounds and destination zip code."

Poor description: "Does shipping stuff"

Choosing Parameters

Define only the parameters your tool truly needs:

What You Get Back

After your custom tool runs:

• Tool Output: Whatever result your tool logic produces
• Success Status: Whether the tool completed successfully
• Execution Info: Details about the tool call

Tips for Success

1. Write clear descriptions - AI decides when to use tools based on descriptions
2. Keep tools focused - One tool, one purpose
3. Use descriptive parameter names - "customer_email" not "ce"
4. Handle edge cases - What if inputs are missing or invalid?
5. Test thoroughly - Verify tools work with various inputs

Troubleshooting

Best Practices

• Name tools clearly - Use action words like "calculate", "validate", "lookup"
• Document thoroughly - Future you will thank present you
• Version your tools - Keep track of changes to tool logic
• Monitor usage - See how often and how well tools are being used
• Iterate and improve - Refine tools based on real usage