All activity
Dedy Ariansyahleft a comment
I don’t think the problem is giving agents sensitive tasks. The problem is we can’t inspect what they actually did. Traditional systems have audit trails. Most AI agents only show the final answer. So the fear isn’t autonomy — it’s opacity. Once agents become traceable and explainable, the conversation changes from “never trust AI” to “trust but verify.” I’ve been experimenting with this idea...
How much do you trust AI agents?
Tracing + evaluation in one open-source tool. LangSmith is closed-source. Langfuse is overcomplicated. Most logging tools lack built-in eval. Auditi combines all three. 2-line auto-instrumentation captures all OpenAI, Anthropic & Google API calls. 7+ LLM-as-Judge evaluators run automatically on traces. Human annotation workflows when AI judges aren't enough. Real-time cost tracking. Turn production traces into fine-tuning datasets. Self-host with docker compose up. Python SDK, FastAPI, React.
AuditiOpen source AI agents observability and evaluation
Dedy Ariansyahleft a comment
Hey Product Hunt! 👋 I'm Dedy, the creator of Auditi. I built this while working on AI agents in production. The hardest part wasn't building the agents — it was knowing if they were actually working well and when/where they are not performing as expected. I tried the existing tools: Langfuse (too complex), LangSmith (closed source, vendor lock-in), and various logging tools (zero evaluation)....
AuditiOpen source AI agents observability and evaluation