rotalabs-comply
AI compliance and audit logging infrastructure with multi-framework support.
Features
- Audit Logging: Encrypted, privacy-preserving audit trails for AI interactions
- 7 Compliance Frameworks: EU AI Act, SOC2, HIPAA, GDPR, NIST AI RMF, ISO 42001, MAS FEAT
- 96 Compliance Rules: Comprehensive coverage across all major AI regulations
- Report Generation: Customizable compliance reports in Markdown, JSON, or HTML
- Privacy-First Design: Hash-only mode or encrypted content storage
- Multiple Storage Backends: File, S3, or in-memory storage
- Async-First: Built for high-performance async applications
Installation
pip install rotalabs-comply
With S3 storage support:
pip install rotalabs-comply[s3]
Quick Start
Audit Logging
import asyncio
from rotalabs_comply import AuditLogger, EncryptionManager, MemoryStorage
async def main():
# Set up encrypted audit logging
encryption = EncryptionManager()
storage = MemoryStorage()
logger = AuditLogger(storage, encryption=encryption, store_content=True)
# Log an AI interaction
entry_id = await logger.log(
input="What is the capital of France?",
output="The capital of France is Paris.",
provider="openai",
model="gpt-4",
safety_passed=True,
latency_ms=245.5,
)
print(f"Logged entry: {entry_id}")
# Retrieve the entry
entry = await logger.get_entry(entry_id)
print(f"Provider: {entry.provider}, Model: {entry.model}")
asyncio.run(main())
Privacy Mode (Hash-Only)
# Only store content hashes, not actual content
logger = AuditLogger(
"/var/log/ai-audit",
store_content=False, # Only store SHA-256 hashes
retention_days=365,
)
Compliance Checking
from rotalabs_comply import EUAIActFramework, SOC2Framework, HIPAAFramework
from rotalabs_comply.frameworks.base import AuditEntry, ComplianceProfile
from datetime import datetime
async def check_compliance():
# Create frameworks
eu_ai = EUAIActFramework()
soc2 = SOC2Framework()
# Create an audit entry to check
entry = AuditEntry(
entry_id="test-001",
timestamp=datetime.utcnow(),
event_type="inference",
actor="user@example.com",
action="Generated text response",
human_oversight=True,
user_notified=True,
)
# Create compliance profile
profile = ComplianceProfile(
profile_id="high-risk",
name="High Risk AI System",
risk_level="high",
)
# Check compliance
result = await eu_ai.check(entry, profile)
print(f"EU AI Act compliant: {result.is_compliant}")
print(f"Violations: {len(result.violations)}")
for violation in result.violations:
print(f" - {violation.rule_id}: {violation.description}")
asyncio.run(check_compliance())
Report Generation
from datetime import datetime, timedelta
from rotalabs_comply import ReportGenerator, MemoryStorage
from rotalabs_comply.core import ComplianceProfile, Framework
async def generate_report():
storage = MemoryStorage()
generator = ReportGenerator(storage)
# Define compliance profile
profile = ComplianceProfile(
frameworks=[Framework.EU_AI_ACT, Framework.SOC2],
risk_level="high",
)
# Generate report for last 30 days
end = datetime.utcnow()
start = end - timedelta(days=30)
report = await generator.generate(
period_start=start,
period_end=end,
profile=profile,
)
# Export to markdown
markdown = generator.export_markdown(report)
print(markdown)
asyncio.run(generate_report())
Compliance Frameworks
7 frameworks with 96 total compliance rules:
| Framework | Description | Rules | Key Categories |
|---|---|---|---|
| EU AI Act | European AI regulation | 8 | transparency, oversight, risk_management |
| SOC2 Type II | AICPA Trust Service Criteria | 10 | security, availability, privacy |
| HIPAA | US healthcare data protection | 8 | access_control, audit, privacy |
| GDPR | EU data protection regulation | 14 | data_protection, consent, data_subject_rights |
| NIST AI RMF | US AI Risk Management Framework | 15 | governance, context, measurement, risk_treatment |
| ISO 42001 | AI Management System standard | 23 | context, leadership, planning, operation |
| MAS FEAT | Singapore financial AI governance | 18 | fairness, ethics, accountability, transparency |
from rotalabs_comply import (
EUAIActFramework,
SOC2Framework,
HIPAAFramework,
GDPRFramework,
NISTAIRMFFramework,
ISO42001Framework,
MASFramework,
)
# Check against multiple frameworks
frameworks = [
EUAIActFramework(),
GDPRFramework(),
MASFramework(),
]
for fw in frameworks:
result = await fw.check(entry, profile)
print(f"{fw.name}: {'PASS' if result.is_compliant else 'FAIL'}")
Storage Backends
File Storage
from rotalabs_comply import AuditLogger, FileStorage
# JSONL files with automatic rotation
storage = FileStorage("/var/log/ai-audit", rotation_size_mb=100)
logger = AuditLogger(storage)
S3 Storage
from rotalabs_comply import AuditLogger, S3Storage
# Requires: pip install rotalabs-comply[s3]
storage = S3Storage(
bucket="my-audit-bucket",
prefix="ai-audit/",
region="us-east-1",
)
logger = AuditLogger(storage)
Memory Storage (Testing)
from rotalabs_comply import AuditLogger, MemoryStorage
storage = MemoryStorage(max_entries=10000)
logger = AuditLogger(storage)
Encryption
All audit content can be encrypted using Fernet symmetric encryption:
from rotalabs_comply import EncryptionManager, generate_key
# Auto-generate key
encryption = EncryptionManager()
key = encryption.get_key() # Save this securely!
# Or provide your own key
key = generate_key()
encryption = EncryptionManager(key=key)
# Use with AuditLogger
logger = AuditLogger(
storage,
encryption=encryption,
store_content=True, # Store encrypted content
)
API Reference
Core Types
RiskLevel- Enum: LOW, MEDIUM, HIGH, CRITICALFramework- Enum: EU_AI_ACT, SOC2, HIPAA, GDPR, NIST_AI_RMF, ISO_42001, MASAuditEntry- Audit log entry data modelComplianceProfile- Compliance configurationComplianceViolation- Detected violationComplianceCheckResult- Framework check result
Audit Module
AuditLogger- Main audit logging interfaceEncryptionManager- Encryption utilitiesFileStorage- JSONL file storageMemoryStorage- In-memory storageS3Storage- AWS S3 storage
Frameworks
EUAIActFramework- EU AI Act compliance (8 rules)SOC2Framework- SOC2 Type II compliance (10 rules)HIPAAFramework- HIPAA compliance (8 rules)GDPRFramework- GDPR compliance (14 rules)NISTAIRMFFramework- NIST AI RMF compliance (15 rules)ISO42001Framework- ISO 42001 compliance (23 rules)MASFramework- MAS FEAT compliance (18 rules)
Reports
ReportGenerator- Generate compliance reportsComplianceReport- Report data modelReportSection- Report section
Links
- Documentation: https://rotalabs.github.io/rotalabs-comply/
- PyPI: https://pypi.org/project/rotalabs-comply/
- GitHub: https://github.com/rotalabs/rotalabs-comply
- Website: https://rotalabs.ai
- Contact: research@rotalabs.ai
License
Apache-2.0 License - see LICENSE for details.
Metadata
Release files for rotalabs-comply 1.1.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| rotalabs_comply-1.1.0.tar.gz | 1.6 MB | Without provenance | Details for rotalabs_comply-1.1.0.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| rotalabs_comply-1.1.0-py3-none-any.whl | Python 3 | none | any | Details for rotalabs_comply-1.1.0-py3-none-any.whl |
Total release size: 1.7 MB
Release files / rotalabs_comply-1.1.0.tar.gz
| Download URL | rotalabs_comply-1.1.0.tar.gz |
|---|---|
| Size | 1.6 MB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
dd419d5327a07b7845b022f3602e84788d67cbed5f25883e84386d6297b6a917
|
|
BLAKE2b-256 checksum How to use checksums |
b698dc74760219b660bbe32ddeac007b8c16da52061ebb3b2bc70e6db6dab6eb
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.12.6
|
Release files / rotalabs_comply-1.1.0-py3-none-any.whl
| Download URL | rotalabs_comply-1.1.0-py3-none-any.whl |
|---|---|
| Size | 111.8 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
060a0f7a19d2703accbeded53332be094ca37961b4675209836966b174d3f746
|
|
BLAKE2b-256 checksum How to use checksums |
131924d70ebf9411d93674a7bcb2a8760ad729c12ae60637c61d1292e8307fa0
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.12.6
|
Security for rotalabs-comply 1.1.0
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.