The Assistant Sandbox is a testing environment where you can experiment with your voice assistant’s system prompt, make real-time tweaks, and trigger web calls — all without modifying your production assistant. It’s the fastest way to fine-tune how your assistant behaves during a call.
Credits are deducted for sandbox calls. Sandbox calls are real calls and
will consume credits from your wallet just like production calls.
Enter your Public Key, Private Key, and Workspace credentials to access the sandbox, and click the Validate button.
Sandbox authentication — enter your Public Key, Private Key, and Workspace ID to validate access.
Field
Description
Public Key
Your public API key
Private Key
Your private/secret API key
Workspace
The workspace identifier you want to work with
Once authenticated, the sandbox loads your workspace data and retrieves all assistants configured under it.
2
Select an Assistant
After authentication, a dropdown appears listing all assistants created in your workspace. Select the assistant you want to test.
Assistant selection — choose from all assistants configured in your workspace.
3
Review the Prompt & Configure
Once you select an assistant, the screen splits into two sections:
Sandbox main interface — system prompt on the left, payload and metadata on the right.
Left Section — System Prompt & Variables
The assistant’s system prompt is displayed in an editable area, organized by sections like Context, Objective, and Personality
Below the prompt, you’ll see allowed variables — these are system-generated variables based on the data you provide (e.g., customer name, order details, discount info)
You can only use these system-generated variables in your prompt
Right Section — Configuration
Payload — A JSON text box where you paste the call payload data (e.g., customer info, cart details)
Metadata — A JSON text box where you paste any additional metadata for the call
Change Assistant — A button in the top-right to switch to a different assistant
4
Make a Call
Populate the Payload and Metadata fields with your JSON data 2.
Optionally, edit the system prompt to test different conversation
behaviors 3. Click the Test Assistant button to trigger a web call 4.
You’ll be connected to the assistant and can speak with it in real-time
5
Iterate & Refine
After each call, review how the assistant responded:
Modify the system prompt based on your findings
Adjust the Payload or Metadata as needed
Click Test Assistant again to test your changes instantly
Repeat this cycle until the assistant behaves as expected.
6
Switch Assistants
To test a different assistant, click the Change Assistant button in the top-right corner. The assistant dropdown will reappear, letting you select another assistant from your workspace.
Prompt changes are not saved. Any modifications you make to the system
prompt in the sandbox are temporary. They are not stored or updated in the
database. To apply changes permanently, update the assistant’s prompt through
the dashboard or API.
Use only system-generated variables. The allowed variables shown in the
sandbox are automatically generated based on the data you provide. Custom or
arbitrary variables will not be resolved during the call.
No events or actions in sandbox. Webhook events (such as call.ended,
call.completed, end-of-call, action, etc.) and actions (such as
create-order, send-whatsapp) are not triggered during sandbox calls.
The sandbox is purely for testing prompt behavior and conversation flow.
Start with your production prompt — Select the assistant and test its current behavior before making changes.
Change one thing at a time — Modify a single aspect of the prompt per call so you can clearly identify what impacts the assistant’s behavior.
Test edge cases — Vary your payload data (e.g., different product types, price ranges, customer profiles) to ensure the prompt handles diverse scenarios.
Note your changes — Since sandbox edits aren’t saved, copy your final working prompt and apply it to your assistant configuration manually.
Monitor your credits — Each sandbox call deducts credits, so plan your testing sessions efficiently.
