indro 1.0.2

Production-ready SDK for the Indro AI Ecosystem

Indro SDK

Production-ready Python SDK for the Indro Edge API.

The Indro SDK provides a secure, high-performance, enterprise-grade client for interacting with the Indro Edge inference platform.

Base API Endpoint:

https://abhinav337463-indro-edge-api.hf.space

---

Features

Core Features

• Fully typed API
• Automatic request serialization
• Automatic response parsing
• GET, POST, PUT, PATCH, DELETE support
• JSON requests
• Form data requests
• Multipart file uploads
• Streaming response support
• Connection pooling
• Keep-alive support
• Batch request support
• Request cancellation
• Request deduplication
• Offline request queue
• Middleware system
• Interceptor system
• Plugin system
• Event hooks
• Fluent API design

---

Security Features

• API Key Authentication
• Bearer Token Authentication
• Custom Header Authentication
• Secure HTTPS enforcement
• SSL validation
• Request signing
• HMAC support
• Header sanitization
• Input sanitization
• Output sanitization
• Secret masking in logs
• Anti-injection protections
• Secure error handling
• Environment variable support

---

Performance Features

• In-memory cache
• Persistent cache
• Smart cache invalidation
• Compression support
• Lazy loading
• Request prioritization
• Concurrent request management
• Queue management
• Optimized memory usage

---

Monitoring Features

• Request metrics
• Performance metrics
• Latency tracking
• Correlation IDs
• Request tracing
• Error tracking
• Health checks
• Telemetry hooks

---

Installation

Install directly:

pip install indro

Development installation:

pip install indro[dev]

WebSocket support:

pip install indro[ws]

Secure token storage support:

pip install indro[keyring]

---

Quick Start

from main import IndroClient

client = IndroClient()

result = client.predict( model_name="house-price-model", features=[ 1200, 3, 2, 5 ] )

print(result.prediction)

---

Initialization

Basic:

from main import IndroClient

client = IndroClient()

Custom endpoint:

client = IndroClient( base_url="https://abhinav337463-indro-edge-api.hf.space" )

Debug mode:

client = IndroClient( debug=True )

Verbose logging:

client = IndroClient( verbose=True )

---

Authentication

API Key

client = IndroClient( api_key="YOUR_API_KEY" )

---

Bearer Token

client = IndroClient( bearer_token="YOUR_TOKEN" )

---

Custom Headers

client = IndroClient( headers={ "X-Custom": "value" } )

---

Available API Methods

List Models

Maps to:

GET /api/models

Example:

models = client.models()

print(models.models)

---

Predict

Maps to:

POST /api/predict

Example:

response = client.predict( model_name="house-price-model", features=[ 1200, 3, 2, 5 ] )

print(response.prediction)

---

Async Predict

response = await client.predict_async( model_name="house-price-model", features=[1,2,3,4] )

---

Batch Requests

responses = client.batch_predict([ { "model_name": "model-a", "features": [1,2,3] }, { "model_name": "model-b", "features": [4,5,6] } ])

---

Health Checks

The SDK includes helpers for operational endpoints.

Health

health = client.health()

Endpoint:

GET /health

---

Readiness

ready = client.ready()

Endpoint:

GET /ready

---

Metrics

metrics = client.metrics()

Endpoint:

GET /metrics

---

SLO Dashboard

slo = client.slo()

Endpoint:

GET /slo

---

Caching

Enable memory cache:

client = IndroClient( cache_enabled=True )

Custom TTL:

client = IndroClient( cache_ttl=300 )

Clear cache:

client.cache.clear()

---

Retry System

Automatic retry support:

client = IndroClient( retries=5 )

Exponential backoff:

client = IndroClient( retry_backoff_factor=2 )

---

Timeout Management

client = IndroClient( timeout=60 )

Per request:

client.predict( model_name="my-model", features=[1,2,3], timeout=120 )

---

Middleware

def middleware(request): print(request.url) return request

client.add_middleware(middleware)

---

Interceptors

Request interceptor:

client.interceptors.request.use( lambda req: req )

Response interceptor:

client.interceptors.response.use( lambda res: res )

---

Event Hooks

@client.on("request") def on_request(event): print(event)

@client.on("response") def on_response(event): print(event)

@client.on("error") def on_error(event): print(event)

---

Error Handling

Base error:

from main import IndroSDKError

Authentication:

from main import AuthenticationError

Validation:

from main import ValidationError

Timeout:

from main import TimeoutError

Network:

from main import NetworkError

Server:

from main import ServerError

Retry exhaustion:

from main import RetryExhaustedError

Example:

try: client.predict( model_name="demo", features=[1,2,3] ) except ValidationError: pass except NetworkError: pass

---

Environment Variables

INDRO_BASE_URL=https://abhinav337463-indro-edge-api.hf.space

INDRO_API_KEY=your_api_key

INDRO_BEARER_TOKEN=your_token

INDRO_TIMEOUT=60

INDRO_DEBUG=false

---

Production Recommendations

For production deployments:

client = IndroClient( timeout=60, retries=5, cache_enabled=True, verify_ssl=True, keep_alive=True, compression=True )

---

API Compatibility

Compatible with:

• Indro Edge API v3.x
• Future v3 minor releases
• Backward compatibility layer included

---

Thread Safety

The SDK is designed to be thread-safe and suitable for:

• Web servers
• FastAPI
• Django
• Flask
• Background workers
• Microservices
• Enterprise applications

---

License

MIT License

Copyright (c) 2026

Abhinav Anand

---

Example

from main import IndroClient

client = IndroClient( debug=True, cache_enabled=True )

models = client.models()

print(models.models)

prediction = client.predict( model_name=models.models[0], features=[1.0, 2.0, 3.0, 4.0] )

print(prediction)

Built for the Indro Edge ecosystem.