Hyperfence Docs

Hyperfence is a transparent proxy that sits between your application and any LLM provider. Every prompt and response is validated, classified, and logged - without calling a second guard LLM.

Quick Start

Connect your LLM with the current provider-specific proxy URLs, then add your token and send requests as normal.

Hyperfences

How vector Hyperfences work, what's included on each plan, and how to train custom rules.

Providers

Major LLM providers are included out of the box: OpenAI, Anthropic, Gemini, DeepSeek, Grok, and more.

Analytics

See token usage, blocked events, topic trends, sentiment, and response quality - all in one place.

How it works

Hyperfence speaks the same API protocols as your LLM provider. Your application changes one value: the base URL. Requests flow through the proxy, are validated by purpose-built ML classifiers, and are forwarded to the upstream model. Validation adds under 200 ms of latency. No guard LLM. No SDK. No code changes beyond that one line.