> ## Documentation Index
> Fetch the complete documentation index at: https://clerkchat.co/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Test your agent before publishing

> Use the Playground to verify answers, models, and conversation starters.

The Playground runs private test conversations against your agent's ready knowledge sources. Use it after changing sources, instructions, or automations.

<Steps>
  <Step title="Open the Playground">
    Select an agent, then select **Playground** in the sidebar.
  </Step>

  <Step title="Choose a model">
    Select the model you want to test and save the selection. Model availability may depend on your
    plan.
  </Step>

  <Step title="Test expected answers">
    Ask direct customer questions covered by your knowledge. Check facts, tone, links, and any
    automation that runs.
  </Step>

  <Step title="Test support boundaries">
    Ask about unsupported topics and requests that require a person. Confirm that the agent follows
    your instructions instead of inventing an answer.
  </Step>

  <Step title="Start a new chat">
    Click **New Chat** before testing a revised setup. This removes the prior conversation context
    from your next test.
  </Step>
</Steps>

## Configure conversation starters

Add all 3 conversation starters in the Playground and save them. Published help pages show these prompts, and widgets can show them when **Show conversation starters** is enabled.

<Note>
  An indexing notice means new website content isn't ready yet. Ready sources remain available
  during indexing.
</Note>


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.