# Adrian

> Adrian is an open-source runtime security tool that monitors AI agent reasoning and actions, spotting injected prompts and unsafe tool calls before they execute.

- **Category:** Security
- **Licence:** Apache-2.0
- **Latest version:** python-sdk-v1.2.1
- **Open source:** oss
- **Languages:** Typescript, Go, Python
- **Platforms:** Docker Compose
- **Tags:** Mcp, Ai Agents, Agents, Observability, Agentic Ai, Anthropic, Openai, Langchain, Next.js, AI, Go, Docker Compose, Threat detection, Self Hosted, Apache 2.0, Claude code, Ai governance, Ai security, Prompt injection, Langgraph, Runtime security, Ai agent security, Llama Cpp, Llm Security, Prompt Injection Detection, Mcp Security, Openai Agents, Runtime Monitoring, Tool Use Control, Agent Security, Aarm, Agentic Security

## Description

Adrian is an open-source runtime security engine for AI agents. It watches both agent activity logs and reasoning traces to spot malicious or out-of-scope behaviour, and can step in before actions run. The tool catches prompt injection, tool poisoning, data exfiltration, and privilege escalation. It provides Python SDKs for LangChain and Anthropic, a TypeScript SDK for OpenAI, and a native Claude Code plugin. Adrian can be self-hosted offline using Docker Compose with a Go backend, Next.js dashboard, and local Llama.cpp container running a Gemma model. Deployment requires an NVIDIA GPU with CUDA. Licensed under Apache 2.0.

## Links

- [Homepage](https://github.com/secureagentics/adrian)
- [Website](https://secureagentics.ai)
- [Documentation](https://docs.adrian.secureagentics.ai)
- [GitHub Repository](https://github.com/secureagentics/Adrian)
- [Discord Community](https://discord.gg/Vq2VyYrw8Z)
- [LinkedIn](https://www.linkedin.com/company/secure-agentics)
- [Product Hunt](https://www.producthunt.com/products/adrian)

---
Source: https://selfhost.directory/project/adrian