> This page is for version Latest (default).
> For other versions, use one of these documentation indexes:
> - Latest (default): https://docs.nvidia.com/nemo/guardrails/latest/llms.txt
> - v0.25.0: https://docs.nvidia.com/nemo/guardrails/v0.25.0/llms.txt
> - v0.24.1: https://docs.nvidia.com/nemo/guardrails/v0.24.1/llms.txt
> - v0.24.0: https://docs.nvidia.com/nemo/guardrails/v0.24.0/llms.txt
> - v0.23.0: https://docs.nvidia.com/nemo/guardrails/v0.23.0/llms.txt
> - v0.22.0: https://docs.nvidia.com/nemo/guardrails/v0.22.0/llms.txt

> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.nvidia.com/nemo/guardrails/llms.txt.

# Use the Guardrails Hooks Server

> Check the tool calls of Claude Code, Codex, and NeMo Relay before they run, with the Guardrails hooks server.

The NVIDIA NeMo Guardrails library includes the Guardrails hooks server (`nemoguardrails hooks-server`), which checks a coding agent's tool calls before the agent runs them.
The agent's own pre-tool hook sends each call to the hooks server, which applies a tool policy and tool-call rails to it and answers with allow or deny.

The hooks server is a separate process from the [Guardrails API server](/nemo/guardrails/latest/run-guardrailed-inference/using-fastapi-server) (`nemoguardrails server`).
It serves only the agent hook endpoints, and it never calls a model to generate a response.

#### [Overview](/nemo/guardrails/latest/run-guardrailed-inference/using-hooks-server/overview)

How the hooks server checks a tool call, how it differs from the Guardrails API server, and a quick start.

Concept

#### [Run the Hooks Server](/nemo/guardrails/latest/run-guardrailed-inference/using-hooks-server/run-hooks-server)

Start the server, write its tool policy and rails, and monitor it with logs, metrics, and traces.

Tutorial

#### [Claude Code](/nemo/guardrails/latest/run-guardrailed-inference/using-hooks-server/claude-code)

Send every Claude Code tool call to the hooks server with an HTTP hook.

How To

#### [Codex](/nemo/guardrails/latest/run-guardrailed-inference/using-hooks-server/codex)

Send every Codex tool call to the hooks server with a command hook.

How To

#### [NeMo Relay Plugin](/nemo/guardrails/latest/run-guardrailed-inference/using-hooks-server/nemo-relay-plugin)

Check every tool call NeMo Relay manages with the NVIDIA NeMo Guardrails Plugin.

How To

#### [NeMo Relay Endpoint](/nemo/guardrails/latest/run-guardrailed-inference/using-hooks-server/nemo-relay)

The endpoint a NeMo Relay plugin's tool guardrail calls.

Reference

#### [Results and Errors](/nemo/guardrails/latest/run-guardrailed-inference/using-hooks-server/results-and-errors)

What each agent path returns for an allow, a deny, and an error, and how to make an agent block a call when a check fails.

Reference

#### [Hooks Server API](/nemo/guardrails/latest/reference/hooks-server-api/overview)

The request and response fields of each hooks server endpoint.

Reference