Feature
Zero Completion Insurance
100% guarantee that your requests complete successfully or your money back
What is Zero Completion Insurance?
Zero Completion Insurance ensures that every request you send to ParrotRouter either completes successfully or you get a full refund. We handle all the complexity of retries, failovers, and persistent queuing so your critical AI operations never fail.
100% Guarantee
Every request completes or full refund
Unlimited Retries
We retry across providers until success
Persistent Queue
Requests survive even system restarts
How It Works
Enable Zero Completion Insurance for mission-critical requests:
Enable Zero Completion Insurancepython
from openai import OpenAI
client = OpenAI(
base_url="https://api.parrotrouter.com/v1",
api_key="your-api-key"
)
# Enable insurance for critical request
response = client.chat.completions.create(
model="gpt-4-turbo-preview",
messages=[{
"role": "user",
"content": "Process critical business logic"
}],
extra_headers={
"X-Zero-Completion-Insurance": "true",
"X-Request-Priority": "critical",
"X-Max-Wait-Time": "3600" # Max 1 hour
}
)
# Check insurance status
print(f"Insurance ID: {response.headers.get('X-Insurance-ID')}")
print(f"Guaranteed by: {response.headers.get('X-Guarantee-Expires')}")
# Track insured request
import requests
status = requests.get(
f"https://api.parrotrouter.com/v1/insurance/{insurance_id}",
headers={"Authorization": "Bearer your-api-key"}
).json()
print(f"Status: {status['status']}") # pending, processing, completed, refunded
print(f"Attempts: {status['attempts']}")
print(f"Providers tried: {status['providers_attempted']}")
print(f"Current queue position: {status['queue_position']}")Zero Completion Insurance adds a small fee (typically $0.001 per request) but guarantees completion.
Insurance Levels
Standard Insurance
Default
Basic protection with standard retry logic.
# Standard insurance configuration
response = client.chat.completions.create(
model="gpt-4-turbo-preview",
messages=[{"role": "user", "content": "Important request"}],
extra_headers={
"X-Zero-Completion-Insurance": "standard",
"X-Insurance-Config": json.dumps({
"max_attempts": 10,
"retry_delay": "exponential",
"max_wait_time": 600, # 10 minutes
"fallback_models": ["gpt-3.5-turbo", "claude-3-haiku"],
"refund_on_timeout": True
})
}
)✓ Up to 10 retry attempts
✓ Automatic provider failover
✓ 10-minute guarantee window
✓ Full refund if failed
Premium Insurance
Recommended
Enhanced protection with priority processing.
# Premium insurance configuration
response = client.chat.completions.create(
model="gpt-4-turbo-preview",
messages=[{"role": "user", "content": "Mission critical operation"}],
extra_headers={
"X-Zero-Completion-Insurance": "premium",
"X-Insurance-Config": json.dumps({
"max_attempts": "unlimited",
"retry_strategy": "aggressive",
"max_wait_time": 3600, # 1 hour
"priority_queue": True,
"multi_region": True,
"preserve_context": True,
"notification_webhook": "https://your-app.com/webhook"
})
}
)✓ Unlimited retry attempts
✓ Priority queue placement
✓ Multi-region failover
✓ 1-hour guarantee window
✓ Real-time status updates
✓ Context preservation
Enterprise Insurance
Custom
Custom SLA with dedicated resources.
# Enterprise insurance configuration
response = client.chat.completions.create(
model="gpt-4-turbo-preview",
messages=[{"role": "user", "content": "Enterprise critical process"}],
extra_headers={
"X-Zero-Completion-Insurance": "enterprise",
"X-Insurance-Config": json.dumps({
"sla_level": "custom",
"dedicated_capacity": True,
"max_wait_time": 7200, # 2 hours
"guaranteed_throughput": 1000, # requests/minute
"private_queue": True,
"audit_trail": True,
"compensation_multiplier": 10 # 10x refund on failure
})
}
)✓ Custom SLA terms
✓ Dedicated infrastructure
✓ Guaranteed throughput
✓ Private priority queue
✓ Complete audit trail
✓ Enhanced compensation
Persistent Queue System
Requests are stored in a durable queue that survives any failure:
Queue Managementpython
# Submit to persistent queue
response = client.chat.completions.create(
model="gpt-4-turbo-preview",
messages=[{"role": "user", "content": "Process when available"}],
extra_headers={
"X-Zero-Completion-Insurance": "true",
"X-Queue-Config": json.dumps({
"queue_type": "persistent",
"priority": "high", # critical, high, normal, low
"dedupe_key": "order-12345", # Prevent duplicates
"batch_id": "batch-001", # Group related requests
"execute_after": "2024-01-15T10:00:00Z", # Scheduled execution
"execute_before": "2024-01-15T12:00:00Z", # Deadline
"idempotency_key": "unique-request-id"
})
}
)
# Monitor queue status
queue_status = requests.get(
"https://api.parrotrouter.com/v1/queue/status",
headers={"Authorization": "Bearer your-api-key"},
params={"batch_id": "batch-001"}
).json()
print(f"Queue depth: {queue_status['total_pending']}")
print(f"Processing rate: {queue_status['requests_per_minute']}")
print(f"Estimated completion: {queue_status['estimated_completion_time']}")
# Bulk queue management
bulk_response = requests.post(
"https://api.parrotrouter.com/v1/queue/bulk",
headers={"Authorization": "Bearer your-api-key"},
json={
"requests": [
{
"model": "gpt-4-turbo-preview",
"messages": [{"role": "user", "content": f"Process item {i}"}],
"insurance": "premium",
"priority": "normal"
}
for i in range(100)
],
"batch_config": {
"process_in_order": True,
"stop_on_failure": False,
"notification_url": "https://your-app.com/batch-complete"
}
}
)Failure Handling
Automatic Recovery
# Configure recovery strategies
recovery_config = {
"strategies": [
{
"error_type": "timeout",
"action": "retry_with_smaller_chunks",
"config": {
"chunk_size": 0.5, # Reduce by 50%
"max_reductions": 3
}
},
{
"error_type": "context_too_long",
"action": "summarize_and_retry",
"config": {
"summarization_model": "gpt-3.5-turbo",
"preserve_key_points": True
}
},
{
"error_type": "rate_limit",
"action": "queue_with_backoff",
"config": {
"initial_delay": 60,
"max_delay": 3600
}
},
{
"error_type": "model_error",
"action": "fallback_cascade",
"config": {
"models": [
"gpt-4-turbo-preview",
"claude-3-opus",
"gpt-3.5-turbo"
]
}
}
]
}
response = client.chat.completions.create(
model="gpt-4-turbo-preview",
messages=long_conversation,
extra_headers={
"X-Zero-Completion-Insurance": "true",
"X-Recovery-Config": json.dumps(recovery_config)
}
)Compensation & Refunds
# Check insurance claims
claims = requests.get(
"https://api.parrotrouter.com/v1/insurance/claims",
headers={"Authorization": "Bearer your-api-key"},
params={"status": "all", "days": 30}
).json()
for claim in claims['claims']:
print(f"
Claim ID: {claim['id']}")
print(f"Request ID: {claim['request_id']}")
print(f"Status: {claim['status']}") # pending, approved, credited
print(f"Failure reason: {claim['failure_reason']}")
print(f"Refund amount: $" + str(claim['refund_amount']))
print(f"Compensation: $" + str(claim['additional_compensation']))
# File a manual claim (rarely needed)
claim = requests.post(
"https://api.parrotrouter.com/v1/insurance/claims",
headers={"Authorization": "Bearer your-api-key"},
json={
"request_id": "req_abc123",
"reason": "Response not received within SLA",
"evidence": {
"expected_completion": "2024-01-15T10:00:00Z",
"actual_completion": None,
"insurance_id": "ins_xyz789"
}
}
)Use Cases
Financial Transactions
Ensure critical financial operations complete.
# Process payment with guarantee
response = client.chat.completions.create(
model="gpt-4-turbo-preview",
messages=[{
"role": "user",
"content": f"Process payment of $1,234.56 for order {order_id}"
}],
extra_headers={
"X-Zero-Completion-Insurance": "premium",
"X-Transaction-ID": transaction_id,
"X-Idempotency-Key": payment_key,
"X-Critical-Operation": "true"
}
)Legal Document Processing
Guarantee processing of time-sensitive documents.
# Process legal filing
response = client.chat.completions.create(
model="gpt-4-turbo-preview",
messages=[{
"role": "user",
"content": "Review and summarize contract"
}],
extra_headers={
"X-Zero-Completion-Insurance": "enterprise",
"X-Deadline": "2024-01-15T17:00:00Z",
"X-Compliance-Mode": "legal",
"X-Audit-Required": "true"
}
)Healthcare Operations
Ensure patient-critical AI operations succeed.
# Process medical analysis
response = client.chat.completions.create(
model="gpt-4-turbo-preview",
messages=[{
"role": "user",
"content": "Analyze patient symptoms"
}],
extra_headers={
"X-Zero-Completion-Insurance": "premium",
"X-Priority": "emergency",
"X-HIPAA-Compliant": "true",
"X-Max-Wait": "300" # 5 minutes max
}
)Batch Processing
Guarantee large batch operations complete fully.
# Process large dataset
for batch in dataset_batches:
response = client.chat.completions.create(
model="gpt-3.5-turbo",
messages=[{
"role": "user",
"content": f"Process batch: {batch['id']}"
}],
extra_headers={
"X-Zero-Completion-Insurance": "standard",
"X-Batch-ID": batch_id,
"X-Batch-Total": len(dataset_batches),
"X-Continue-On-Failure": "false"
}
)Insurance Analytics
Monitor your insurance usage and claims:
Insurance Metricspython
# Get insurance analytics
analytics = requests.get(
"https://api.parrotrouter.com/v1/insurance/analytics",
headers={"Authorization": "Bearer your-api-key"},
params={"period": "month"}
).json()
print(f"Total insured requests: {analytics['total_insured']}")
print(f"Success rate: {analytics['success_rate']}%")
print(f"Average retry count: {analytics['avg_retries']}")
print(f"Total claims: {analytics['total_claims']}")
print(f"Claims approved: {analytics['claims_approved']}")
print(f"Total refunds: $" + str(analytics['total_refunds']))
print(f"Insurance ROI: {analytics['roi']}x")
# Failure analysis
print("
Top failure reasons:")
for reason in analytics['failure_reasons']:
print(f"- {reason['type']}: {reason['count']} ({reason['percentage']}%)")
# Cost benefit analysis
print(f"
Cost Analysis:")
print(f"Insurance fees paid: $" + str(analytics['insurance_fees']))
print(f"Potential losses prevented: $" + str(analytics['losses_prevented']))
print(f"Net benefit: $" + str(analytics['net_benefit']))Best Practices
- 1.Use for Critical Operations Only
Insurance adds overhead - use it for important requests
- 2.Set Realistic Timeouts
Longer timeouts increase success rate but may delay refunds
- 3.Use Idempotency Keys
Prevent duplicate processing of critical requests
- 4.Monitor Insurance Metrics
Track success rates to optimize your configuration