Show HN: Torrix, self hosted, LLM Observability,(no Postgres, no Redis) (github.com)

Torrix, a newly launched self-hosted observability tool for large language models (LLMs), allows users to track essential metrics like tokens, costs, latency, and prompt traces without relying on external databases like Postgres or Redis. It seamlessly integrates with a wide array of LLMs including OpenAI, Anthropic, and Google Gemini, and operates locally using only Docker Desktop. Users can simply run a few commands to set it up and monitor their LLM requests via a user-friendly dashboard. The significance of Torrix lies in its capacity to enhance transparency and control over LLM interactions, particularly crucial for developers and organizations concerned about data privacy and operational efficiency. By enabling comprehensive logging and performance tracking, it helps teams better understand resource usage and optimize LLM performance. Furthermore, the tool's compatibility with various AI services, and its zero-dependency requirement, position it as a valuable asset for the AI/ML community. As organizations increasingly adopt LLMs for diverse applications, tools like Torrix play a vital role in ensuring accountability and resource management across AI workloads.