Vendor-neutral, OpenTelemetry-native AI agent observability, evaluation, governance and runtime control.
Traccia is a vendor-neutral, framework-agnostic control plane for AI agents. Built on OpenTelemetry, it gives engineering and platform teams end-to-end visibility into agent execution, LLM calls, tool usage, cost and evaluation while adding governance and runtime controls. Traccia works across AI models, agent frameworks and cloud environments, helping teams avoid vendor lock-in. It supports integrations including LangChain, CrewAI, OpenAI Agents SDK and Claude Code, with Python and TypeScript SDKs. Teams can trace and debug agents, attribute token and LLM costs, manage agent inventory, detect policy and governance issues, and build an auditable record of agent activity. The goal is simple: help enterprises move from simply observing AI agents to safely operating and controlling them in production.
Live position — it moves as people vote. Awards are granted once the week ends.
Add this to your site or README. It shows your live OSSDrop standing and links back — no account needed.
<a href="https://ossdrop.com/tool/traccia"><img src="https://ossdrop.com/badge/traccia" width="250" height="56" alt="Traccia on OSSDrop" /></a>[](https://ossdrop.com/tool/traccia)What is Traccia?
Traccia is an open-source ai & coding agents tool — Vendor-neutral, OpenTelemetry-native AI agent observability, evaluation, governance and runtime control. It is written in Python.
Is Traccia free and open source?
Yes. Traccia is free and open source, released under the Apache-2.0 license. You can view and contribute to the source at https://github.com/traccia-ai/traccia-py.
What language is Traccia written in?
Traccia is primarily written in Python.
What are open-source alternatives to Traccia?
Other open-source ai & coding agents tools on OSSDrop include Ollama, Transformers, LangChain, llama.cpp.
We built Traccia after seeing a gap between how quickly AI agents are being deployed and how little infrastructure exists to understand and control them once they're in production. We started with an open-source, OpenTelemetry-native SDK because we believe agent infrastructure should be vendor-neutral and framework-agnostic. We're building toward a control plane that helps teams observe, evaluate, govern and ultimately control their AI agents.
Sign in to comment.
This is so easy to use. I have used langsmith and langfuse. But Traccia's policy enforcement is something to looks for.
Good work done folks, loved it!