Skip to main content

Overview

Presets automatically configure the right detection types for your use case. Instead of manually selecting sandbox types, specify a preset and Continum will enable the appropriate checks.

Available Presets

customer-support

For customer support bots, help desks, and Q&A systems.
Automatically enables:
  • PII Detection
  • Content Policy
  • Bias Detection
  • Hallucination Detection
Best for: Chat support, help desk automation, FAQ bots
For legal research, document analysis, and compliance tools.
Automatically enables:
  • PII Detection
  • Legal Compliance
  • Hallucination Detection
  • Bias Detection
  • Citation Verification
Best for: Legal research, contract analysis, compliance tools

fintech-ai

For financial services, trading, and investment advice.
Automatically enables:
  • PII Detection
  • Financial Compliance
  • Security Audit
  • Bias Detection
  • Hallucination Detection
Best for: Trading bots, financial advisors, investment analysis

healthcare-ai

For medical advice, diagnosis, and healthcare applications.
Automatically enables:
  • PII Detection (including health data)
  • Content Policy
  • Bias Detection
  • Hallucination Detection
  • Medical Compliance
Best for: Medical chatbots, diagnosis tools, health records

coding-assistant

For code generation, review, and development tools.
Automatically enables:
  • Security Audit
  • Supply Chain Integrity
  • Prompt Injection
  • Secret Leak Detection
Best for: Code generation, code review, IDE assistants

agent

For AI agents with tool use and autonomous actions.
Automatically enables:
  • Agent Safety
  • Prompt Injection
  • Data Exfiltration
  • Security Audit
  • Scope Creep Detection
Best for: Autonomous agents, tool-using AI, multi-step workflows

content-generation

For writing, marketing, and creative content.
Automatically enables:
  • Content Policy
  • Bias Detection
  • Hallucination Detection
  • Copyright Compliance
Best for: Content writing, marketing copy, creative tools

internal-tool

For internal team productivity and automation.
Automatically enables:
  • PII Detection
  • Security Audit
  • Data Exfiltration
Best for: Internal chatbots, productivity tools, automation

data-pipeline

For ETL, data processing, and analysis.
Automatically enables:
  • PII Detection
  • Data Exfiltration
  • Security Audit
Best for: Data processing, ETL pipelines, analytics

education-ai

For tutoring, learning, and training applications.
Automatically enables:
  • Content Policy
  • Bias Detection
  • Hallucination Detection
  • Age-Appropriate Content
Best for: Tutoring bots, learning platforms, training tools

Compliance Frameworks

Specify compliance frameworks to automatically enable required checks:

Supported Frameworks

Combining Presets and Frameworks

Presets and frameworks work together to provide comprehensive coverage:
This configuration will enable:
  • All checks from fintech-ai preset
  • Additional checks required by FINRA and SOC2

Custom Detection Types

For advanced use cases, you can specify detection types directly:
Available detection types:
  • PII_DETECTION - Personal information
  • BIAS_DETECTION - Discrimination and bias
  • SECURITY_AUDIT - Code injection, vulnerabilities
  • PROMPT_INJECTION - Jailbreaks, goal hijacking
  • DATA_EXFILTRATION - Data leaks, cross-user leaks
  • AGENT_SAFETY - Irreversible actions, scope creep
  • HALLUCINATION_DETECTION - False information
  • CONTENT_POLICY - Harmful content
  • FINANCIAL_COMPLIANCE - Financial regulations
  • LEGAL_COMPLIANCE - Legal requirements
  • MULTI_TURN_ATTACK - Multi-step attacks
  • SUPPLY_CHAIN_INTEGRITY - Dependency security
  • FULL_SPECTRUM - All detection types
  • CUSTOM - Custom rules

Environment-Specific Configuration

Use different presets for different environments:

Best Practices

1. Start with a Preset

Always start with a preset that matches your use case:

2. Add Compliance Frameworks

Specify all applicable frameworks:

3. Use Environment Variables

Keep configuration flexible:

4. Use Environment Variables

Keep configuration flexible:

Next Steps

Alerts

Set up real-time alerts for violations

Blocking Mode

Use blocking mode for high-risk scenarios

Violation Handlers

React to specific violations in code

Configuration

Advanced SDK configuration