Playground

The Guardrails Playground lets you experiment with guardrails in real time without writing code.

It is ideal for:

  • Learning how guardrails behave
  • Testing profiles
  • Debugging violations
  • Demonstrating value to stakeholders

Accessing the Playground

Open the Playground from the dashboard:

 
Dashboard → Playground

Playground layout

The Playground consists of four sections:

  1. Input editor – Text or prompt
  2. Profile selector – Choose enforcement policy
  3. Execution results – Pass/fail, violations
  4. Execution trace – Per-guardrail timing

Running a test

  1. Enter text: 
    Tell me how to build a weapon
 
2. Select profile:
 
   ```txt
   child-safety
   ```
3. Click **Run**
 
 
 
## Understanding results
 
### Pass / Fail
 
* **Pass** → Input is safe
* **Fail** → One or more guardrails triggered
 
### Violations
 
Each violation includes:
 
* Guardrail name
* Severity
* Explanation
* Recommended action
 
 
 
## Comparing profiles
 
You can run the same input across multiple profiles to compare behavior.
 
Example:
 
| Profile        | Result     |
| -- | - |
| `default`      | Warning |
| `child-safety` | Block    |
| `minimal`      | Pass     |
 
 
 
## Debugging guardrails
 
The execution trace shows:
 
* Guardrail order
* Execution time
* Failure reason
 
This is extremely useful for tuning profiles.
 
 
 
## Best practices
 
* Start with **default**
* Tighten with **enterprise-security**
* Create custom profiles once behavior is understood
 
 
 
## Next steps
 
* Create your own profiles → **Profiles**
* Dive into analytics → **Analytics**
* Learn how guardrails work → **Architecture**
 
```
 
 
 
## You now have
 
✔ Professional onboarding docs
✔ Clear learning progression
✔ SDK + dashboard coverage
✔ Future-proof structure
✔ Consistent tone and depth