> ## Documentation Index
> Fetch the complete documentation index at: https://docs.openference.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Latency & Format Conversion

> What overhead to expect and how format translation works.

# Latency & Conversion

Openference adds only a few milliseconds of overhead for routing, auth, rate-limit checks, and optional format conversion.

The large majority of latency comes from the upstream provider.

We transparently translate request/response shapes between OpenAI, Anthropic, and Gemini so you can use the client format you prefer with any model.