# Semantic Firewall

> Security and Data Protection Platform for Agentic Apps & Model harnesses

## Docs

- [Overview](https://dashboard.sfw.stanwith.me/docs/raw/overview): What the Semantic Firewall is: a security and data-protection layer for agentic apps and model harnesses.
- [Getting started](https://dashboard.sfw.stanwith.me/docs/raw/getting-started): How to send traffic through the Semantic Firewall: direct API, transparent proxy, and agent CLI hooks.
- [Transparent proxy](https://dashboard.sfw.stanwith.me/docs/raw/guides/transparent-proxy): Point your OpenAI or Anthropic SDK at the firewall for zero-code-change PII sanitization and injection defense.
- [Agent CLI & hooks](https://dashboard.sfw.stanwith.me/docs/raw/guides/agent-hooks): Guard Claude Code, Codex, or Gemini CLI with pre/post tool-call inspection using the dependency-free sfw client.
- [Direct API](https://dashboard.sfw.stanwith.me/docs/raw/guides/direct-api): Drive the /v1/inspect lifecycle from your own code: drift baselines, sessions, and the PII de-anonymization round-trip.
- [Detect a compromised agent](https://dashboard.sfw.stanwith.me/docs/raw/guides/detect-compromised-agent): Screen model output for signs an injection succeeded, and gate high-risk agent actions with the action gate.
- [REST API reference](https://dashboard.sfw.stanwith.me/docs/raw/api-reference/rest): Full reference for the Semantic Firewall REST API.
- [gRPC API reference](https://dashboard.sfw.stanwith.me/docs/raw/api-reference/grpc): Full reference for the Semantic Firewall gRPC API.