Traccia Documentation
Trace, monitor, and govern your AI agents in production.
Traccia is an OpenTelemetry-based tracing SDK for AI agents and LLM applications. Instrument your agents with a single decorator and gain full visibility into their execution, costs, and behavior across Python and TypeScript.
pip install tracciaChoose Your Path
How It Works
Traccia sits between your application and observability backends, enriching traces with AI-specific metadata:
Quick Example
Add tracing to any function with the @observe() decorator:
from traccia import observe, init
# Initialize Tracciainit()
@observe()def my_agent(query: str) -> str: """Your agent logic here.""" response = call_llm(query) return response
# All LLM calls are automatically tracedresult = my_agent("What is the weather today?")What You Get
Full Trace Visibility
See every step of agent execution, from prompts to tool calls.
Cost Attribution
Track token usage and costs per agent, model, and tool.
Policy Guardrails
Set spend limits, retry caps, and custom rules.
Auto-Instrumentation
Automatic tracing for OpenAI, Anthropic, and FastAPI.
OpenTelemetry Native
Export to any OTLP-compatible backend.
Team Collaboration
Org-level visibility with role-based access.
Governance Hub
AI system registry, reviews, incidents, evidence exports.
PII redaction (SDK)
Optional pattern masking before export (redact_pii).
Next Steps
Learn Core Concepts
Understand traces, spans, and how Traccia works.
Auto-Instrumentation
See which libraries are automatically traced.
Claude Code
Full observability for Claude Code — sessions, tools, and cost, with no code changes.
Framework Integrations
Integrate Traccia with LangChain and OpenAI Agents SDK.
Governance Hub
Registry, reviews, incidents, and compliance evidence.
Set Up Local Tracing
Run Jaeger locally and view your first traces.
© 2026 Traccia.