Hub/Profiles/Default Production
Beta / Experimental

Default Production

Baseline safety and security for most production LLM applications.

#input
#security
2.2k
Views
540
Likes
140
Used

Contributors

GD
AI

Overview

The Default Production profile is our recommended starting point for most commercial LLM applications. It strikes a carefully tuned balance between security, safety, and user experience, ensuring that your application remains helpful while being protected against common threats.

This profile is designed to be "set and forget" for standard use cases like customer support, content generation, and knowledge retrieval. It addresses the top OWASP for LLM vulnerabilities, including prompt injection and PII leakage, without requiring deep security expertise to configure.

Included Guardrails

6 Rules

Key Benefits

Core Security Coverage

Protects against the most prevalent attacks like prompt injection and jailbreaking.

PII Protection

Automatically detects and redact sensitive personal information in both input and output.

Low Latency

Optimized rule set ensures minimal impact on your application response times.

Wait, when should I use this?

General purpose customer support chatbots
Marketing content generation tools
Internal knowledge base Q&A systems

Integration

json
config.json
{
  "profile": "default",
  "config": {
    "pii_threshold": "0.8",
    "block_on_injection": true
  }
}

Frequently Asked Questions

Is this profile suitable for high-risk applications?

For high-risk domains (finance, healthcare), we recommend using their specific profiles which have stricter compliance controls.

Does this affect model creativity?

Minimal impact. The guardrails focus on safety boundaries rather than constraining the style or creativity of the output.