Privacy and Logging
Enterprise-grade privacy controls and comprehensive logging for compliance
Privacy First Architecture
ParrotRouter is designed with privacy at its core. We never store or train on your data, and provide granular controls for enterprise compliance requirements.
Zero Data Retention
Your prompts and completions are never stored by default
End-to-End Encryption
All data in transit encrypted with TLS 1.3
No Training on Your Data
Your data is never used to train AI models
Logging Options
Control exactly what gets logged for debugging, compliance, and analytics:
from openai import OpenAI
client = OpenAI(
base_url="https://api.parrotrouter.com/v1",
api_key="your-api-key"
)
# Minimal logging - only errors
response = client.chat.completions.create(
model="gpt-4-turbo-preview",
messages=[{"role": "user", "content": "Hello"}],
extra_headers={
"X-Logging-Level": "error",
"X-Log-Request": "false",
"X-Log-Response": "false"
}
)
# Full logging for debugging
response = client.chat.completions.create(
model="gpt-4-turbo-preview",
messages=[{"role": "user", "content": "Debug this"}],
extra_headers={
"X-Logging-Level": "debug",
"X-Log-Request": "true",
"X-Log-Response": "true",
"X-Log-Metadata": "true"
}
)
# Compliance logging - metadata only
response = client.chat.completions.create(
model="gpt-4-turbo-preview",
messages=[{"role": "user", "content": "Process payment"}],
extra_headers={
"X-Logging-Level": "info",
"X-Log-Request": "false",
"X-Log-Response": "false",
"X-Log-Metadata": "true",
"X-Log-User-ID": "user-123",
"X-Log-Session-ID": "session-456"
}
)Logging Levels
Only logs errors and failures. No request/response content is stored.
{
"timestamp": "2024-01-15T10:30:00Z",
"level": "error",
"request_id": "req_abc123",
"error": "Model timeout",
"status_code": 504,
"model": "gpt-4-turbo-preview"
}Logs request metadata without content. Useful for analytics and billing.
{
"timestamp": "2024-01-15T10:30:00Z",
"level": "info",
"request_id": "req_abc123",
"model": "gpt-4-turbo-preview",
"tokens": {
"prompt": 150,
"completion": 200,
"total": 350
},
"latency_ms": 1234,
"user_id": "user-123",
"status_code": 200
}Complete logging including request/response content. Use only for debugging.
{
"timestamp": "2024-01-15T10:30:00Z",
"level": "debug",
"request_id": "req_abc123",
"request": {
"model": "gpt-4-turbo-preview",
"messages": [
{"role": "user", "content": "Hello"}
],
"temperature": 0.7
},
"response": {
"choices": [{
"message": {
"role": "assistant",
"content": "Hello! How can I help you today?"
}
}]
},
"metadata": {
"tokens": {"prompt": 10, "completion": 15},
"latency_ms": 456
}
}Data Retention Policies
Standard Tier
- •Error logs: 30 days retention
- •Metadata logs: 90 days retention
- •No content logging by default
- •Automatic deletion after retention period
Enterprise Tier
- •Customizable retention periods
- •Bring your own storage (S3, GCS)
- •Custom encryption keys
- •Audit logs with immutability
Compliance Features
GDPR Compliance
Full support for GDPR requirements including data portability and right to deletion.
# Request data deletion
import requests
response = requests.post(
"https://api.parrotrouter.com/v1/privacy/delete",
headers={
"Authorization": "Bearer your-api-key",
"Content-Type": "application/json"
},
json={
"user_id": "user-123",
"delete_type": "all", # or "logs_only"
"reason": "user_requested"
}
)
# Export user data
response = requests.post(
"https://api.parrotrouter.com/v1/privacy/export",
headers={
"Authorization": "Bearer your-api-key",
"Content-Type": "application/json"
},
json={
"user_id": "user-123",
"format": "json", # or "csv"
"include_logs": True
}
)PCI DSS Support
Automatic PII detection and masking for payment card data.
# Enable PII masking
response = client.chat.completions.create(
model="gpt-4-turbo-preview",
messages=[{
"role": "user",
"content": "Process payment for card 4111-1111-1111-1111"
}],
extra_headers={
"X-PII-Detection": "true",
"X-PII-Masking": "true",
"X-PII-Types": "credit_card,ssn,email"
}
)
# Logged as: "Process payment for card ****-****-****-1111"HIPAA Compliance
Healthcare data protection with audit trails and encryption.
# HIPAA-compliant request
response = client.chat.completions.create(
model="gpt-4-turbo-preview",
messages=[{
"role": "user",
"content": "Summarize patient symptoms"
}],
extra_headers={
"X-Compliance-Mode": "HIPAA",
"X-Audit-User": "dr-smith-123",
"X-Audit-Reason": "patient-consultation",
"X-Encryption": "AES-256-GCM"
}
)Best Practices
- 1.Minimize Logging
Only log what's necessary for your use case
- 2.Use Metadata Instead of Content
Track requests with IDs and metadata rather than full content
- 3.Implement Client-Side Encryption
Encrypt sensitive data before sending to any API
- 4.Regular Audit Reviews
Periodically review logs and access patterns