alura-ai 0.1.1

AI usage tracking and billing SDK

Alura AI Python SDK

Track AI usage and costs with automatic billing integration.

Installation

pip install alura-ai

With OpenAI support:

pip install alura-ai[openai]

Quick Start

Automatic OpenAI Tracking

from alura import Alura, AluraOpenAI
from openai import OpenAI

# Initialize clients
alura = Alura(api_key="your-alura-api-key")
openai_client = OpenAI(api_key="your-openai-key")
alura_openai = AluraOpenAI(openai_client, alura)

# All calls within trace() are automatically tracked
with alura.trace(customer_id="cust-123", agent_id="my-chatbot"):
    response = alura_openai.chat.completions.create(
        model="gpt-4",
        messages=[{"role": "user", "content": "Hello!"}]
    )

Manual Signal Tracking

from alura import Alura

alura = Alura(api_key="your-alura-api-key")

# Track any event
alura.signal(
    event_name="meeting_booked",
    agent_id="sales-agent",
    customer_id="cust-123",
    data={
        "meeting_type": "demo",
        "duration_minutes": 30
    }
)

Bulk Signal Recording

from alura import Alura, Signal

alura = Alura(api_key="your-alura-api-key")

signals = [
    Signal(
        event_name="email_sent",
        agent_id="outreach-agent",
        customer_id="cust-123",
        data={"recipient": "[e-mail]"}
    ),
    Signal(
        event_name="call_made",
        agent_id="outreach-agent",
        customer_id="cust-123",
        data={"duration_seconds": 120}
    ),
]

alura.signal_bulk(signals)

API Reference

Alura(api_key, base_url)

Main client for Alura API.

alura.trace(customer_id, agent_id)

Context manager for tracing calls. All AluraOpenAI calls within are auto-tagged.

alura.signal(event_name, agent_id, data, customer_id)

Record a single signal/event.

alura.signal_bulk(signals)

Record multiple signals in one request.

AluraOpenAI(openai_client, alura_client)

OpenAI wrapper with automatic tracking.

License

MIT License