Noveum Trace - Lightweight Tracing SDK

Noveum Trace is Noveum's lightweight Python SDK for tracing AI and voice applications. It provides automated tracing for LangChain and LangGraph-based applications, along with context manager-based APIs for custom instrumentation.

What is Noveum Trace?

Noveum Trace automatically captures comprehensive traces of your AI application's behavior, including LLM calls, agent interactions, RAG operations, and custom business logic. With minimal overhead and zero configuration required, it's designed for developers who want to add observability to their AI applications quickly.

Key Features

Quick Start

pip install noveum-trace

import noveum_trace
from noveum_trace.context_managers import trace_llm
import openai
 
# Initialize once at application startup
noveum_trace.init(
    api_key=os.getenv("NOVEUM_API_KEY"),
    project="my-ai-app",
    environment="production"
)
 
# Trace LLM calls
with trace_llm(model="gpt-4", provider="openai") as span:
    response = openai.chat.completions.create(
        model="gpt-4",
        messages=[{"role": "user", "content": "Hello"}]
    )
    span.set_attributes({
        "llm.response": response.choices[0].message.content,
        "llm.completion_tokens": response.usage.completion_tokens,
        "llm.prompt_tokens": response.usage.prompt_tokens,
    })

What Gets Traced

• LLM Operations: Model calls, token usage, costs, latency
• RAG Pipelines: Document retrieval, embeddings, context assembly
• Agent Workflows: Multi-agent interactions, tool usage, decision trees
• Voice Agents: STT/TTS operations with LiveKit integration
• Custom Operations: Business logic, external APIs, data processing

Next Steps

• SDK Integration Guide - Complete integration guide with all approaches
• Quick Setup - Get started in 5 minutes
• Integration Examples - See Noveum Trace in action
• Platform SDK Guide - Detailed SDK documentation

Ready to start tracing? Check out the SDK Integration Guide for complete integration instructions!