# Torrix (torrix.ai)

Self-hosted proxy and SDK that logs LLM calls, tracing latency, errors and per-token cost across OpenAI, Anthropic and other providers.

- Atlas card: https://inite.ai/en/atlas/torrix.ai
- Site: https://torrix.ai
- For: AI teams who want self-hosted tracking of LLM costs, latency and failures across providers
- Categories: dev
- Use 12/100 · alive 88/100

## Used for

- Trace each LLM call's tokens, cost and latency
- Track spend by provider and model
- Compare model outputs against saved test datasets

## Price

Not read yet.

## For developers

None found.
- Read 2026-10-19

## Your data

- Certifications stated: none
- Read 2026-10-19

---

Source: INITE Atlas, https://inite.ai/en/atlas/torrix.ai. JSON: https://inite.ai/api/atlas/torrix.ai. MCP: https://inite.ai/api/mcp (atlas_lookup).