Skip to main content
The Tools Tab is where you manage the functions and integrations available to the agent during a call. Tools extend what the agent can do: fetch live data, write records, trigger actions in external systems, or execute any custom logic reachable via HTTP.

​
What Tools Are

Tools in DialNexa are LLM-callable functions. During a call, when the agent determines that a tool call is appropriate, it generates a structured call (function name + arguments). DialNexa executes the call and returns the result to the LLM, which incorporates it into the next response. Tools can be:
  • Custom HTTP functions: Endpoints you define that DialNexa calls with a JSON payload and receives a JSON response from.
  • Pre-built integrations: Connectors to third-party services configured through the DialNexa integrations layer.
  • Built-in functions: Platform-level functions such as call transfer, SMS send, or call end.

​
View Attached Tools

The Tools Tab lists all tools currently attached to this agent. For each tool, you can see:
  • Name: The function name the LLM uses to invoke it.
  • Type: Custom function, integration, or built-in.
  • Status: Enabled or disabled.
  • Last Used: The last time this tool was called on a live call.

​
Add a Custom Function

1

Click Add Tool

Click Add Tool in the Tools Tab. A panel opens with tool configuration options.
2

Select tool type

Select Custom Function to define an HTTP endpoint.
3

Configure the function

Fill in the function definition:
FieldDescription
Function NameThe name the LLM uses to call this function. Use snake_case (e.g., lookup_order, create_ticket). The LLM sees this name in its tool list.
DescriptionA clear description of what this function does. The LLM uses this to decide when to call it. Write it from the LLM’s perspective: “Call this function when the caller asks for their order status and provides an order number.”
Endpoint URLThe HTTPS URL DialNexa calls when the function is triggered. Must be HTTPS.
MethodHTTP method: GET or POST.
HeadersStatic HTTP headers to include with every request (e.g., Authorization: Bearer <token>).
ParametersThe arguments the LLM can pass to this function. Define each parameter with a name, type (string, number, boolean, enum), description, and whether it is required.
4

Save

Click Save. The function is added to the Tools Tab and becomes available to the LLM immediately (for the current draft).

​
Add a Pre-Built Integration

1

Click Add Tool

Click Add Tool, then select Integration.
2

Select the integration

Choose from the available pre-built integrations (e.g., CRM lookup, calendar booking, ticketing system). If the integration is not yet connected, you are prompted to authenticate it first.
3

Configure the integration function

Each integration exposes one or more functions. Select which functions to attach to this agent. Configure any function-level settings (e.g., which CRM fields to return, which calendar to check).
4

Save

Click Save. The integration functions appear in the Tools Tab.

​
Enable and Disable Tools

Toggle the Enabled switch next to any tool to enable or disable it without removing it. Disabled tools are not shown in the LLM’s tool list and will not be called during a call. Use this to:
  • Temporarily disable a tool while its backend is under maintenance.
  • Disable tools that are not relevant for a specific agent version without deleting the configuration.
  • Test agent behavior without a specific tool active.
Enabling or disabling a tool on the Tools Tab requires saving and publishing a new agent version before it takes effect on live calls.

​
Remove a Tool

To permanently remove a tool from the agent:
  1. Find the tool in the list.
  2. Click the Remove icon (trash icon or ”…” menu > Remove).
  3. Confirm the removal.
Removing a tool deletes its configuration from this agent. If you use the same function in multiple agents, removing it here does not affect other agents.

​
Function Descriptions Matter

The LLM decides which tool to call based on the function name, description, and the current conversation context. Ambiguous or generic descriptions lead to incorrect tool selection. Less effective: 
Name: get_data
Description: Gets data from the system.
More effective: 
Name: lookup_account
Description: Call this function when the caller provides their account number or email address and wants to check their account status, balance, or recent activity. Do not call this function unless the caller has provided an identifier.

​
Tool Call Visibility in Testing

In the LLM Playground and simulation testing, tool calls are displayed inline in the conversation thread showing:
  • Function name
  • Arguments passed
  • Response returned (mocked or live)
Use the LLM Playground to verify that the LLM calls the right function with the expected arguments before deploying to production. See LLM Playground.

​
Function Execution Timeout

Each function call has a default execution timeout of 5 seconds. If your endpoint does not respond within this window, DialNexa returns an error to the LLM. The LLM handles this as a tool error and typically informs the caller that the information is temporarily unavailable. If your endpoint consistently takes longer than 5 seconds, optimize the endpoint or contact DialNexa support to discuss timeout configuration options.

​
Security

  • All function endpoints must use HTTPS.
  • Use the Headers field to include authentication tokens (e.g., Authorization: Bearer <token>). Headers are encrypted at rest.
  • Do not include secrets in function parameter definitions. Parameters are included in LLM context and may appear in logs.
  • DialNexa signs outbound function requests with a webhook signature header. Verify this signature in your endpoint to confirm requests originate from DialNexa.