aetherlab 0.3.1

Official Python SDK for AetherLab's AI Guardrails and Compliance Platform

AetherLab Python SDK

Official Python SDK for the AetherLab AI Control Layer.

Installation

pip install aetherlab

Quick Start

from aetherlab import AetherLabClient

# Initialize the client
client = AetherLabClient(api_key="your-api-key")

# Validate content (recommended new API)
result = client.validate_content(
    content="Your AI-generated content here",
    content_type="customer_support",
    desired_attributes=["helpful", "professional"],
    prohibited_attributes=["rude", "misleading"]
)

# Check compliance metrics
print(f"Compliant: {result.is_compliant}")
print(f"Probability of non-compliance: {result.avg_threat_level:.1%}")
print(f"Confidence in compliance: {result.confidence_score:.1%}")

if result.is_compliant:
    print(f"✅ Content is safe: {result.content}")
else:
    print(f"❌ Issues found: {result.violations}")
    print(f"💡 Suggestion: {result.suggested_revision}")

# Legacy API (still supported)
result = client.test_prompt(
    user_prompt="Hello, how can I help?",
    blacklisted_keywords=["harmful", "dangerous"]
)
print(f"Compliant: {result.is_compliant}")

Features

• ✅ Content validation with context-aware analysis
• ✅ Multi-language support
• ✅ Real-time compliance checking
• ✅ Suggested revisions for non-compliant content
• ✅ Legacy API compatibility
• ✅ Media analysis capabilities
• ✅ Audit logging

Documentation

For full documentation, visit docs.aetherlab.ai

Examples

See the examples directory for complete examples.

Support

• GitHub Issues: github.com/AetherLabCo/aetherlab-community/issues
• Email: [e-mail]

License

MIT