#55th of 9 in Security
guardrails
Programmable input, output and dialog guardrails for LLM applications
- Stars
- 7.3k
- License
- custom license
- Last commit
- Oct 2026
- Last release
- Sep 2026
- Language
- Python
Missed the latest GitHub check; these facts may be older.
Overview
Python library that sits between application code and the LLM and applies input, dialog, retrieval, execution and output rails. Rails are configured in config.yml and Colang files, and run through the LLMRails API or an HTTP server with a /v1/chat/completions endpoint. Built-in rails cover jailbreak detection, sensitive data masking, fact-checking, hallucination checks and NVIDIA content and topic safety models.
Who it is for: Developers adding safety and topic controls to LLM chat apps and RAG
Strengths
- Five rail types: input, dialog, retrieval, execution and output
- Built-in rails for jailbreak, PII masking, fact-checking and hallucination checks
- Python API plus CLI server with OpenAI-style chat completions endpoint
- Includes an evaluation command and LLM vulnerability scan reports
Weaknesses
- Dialog rails require learning Colang, a separate modeling language
- Sends anonymous telemetry to NVIDIA by default; opt-out needed
- LangChain integration is opt-in via environment variable
- License is flagged NOASSERTION by detection while README badge says Apache 2.0
What it needs
- Docker
- Needs Python 3.10-3.13, LLM provider such as OpenAI or NIM
- Models: OpenAI GPT-3.5/GPT-4, LLaMa-2, Falcon, Vicuna, Mosaic
Also in Security
See all 9| Rank | Project | Score |
|---|---|---|
| 1 | shannonAutonomous pentester that exploits web apps and APIs using source code analysis | 79 out of 100 |
| 2 | StrixAutonomous AI pentesting agents that validate vulnerabilities with working exploits | 75 out of 100 |
| 3 | pentagiAutonomous multi-agent penetration testing system with sandboxed security tools | 75 out of 100 |
| #4 | ifixaiPython CLI that audits AI agents for policy, access and reliability failures | 65 out of 100 |
| #6 | garakCommand-line scanner that probes LLMs for jailbreaks, leakage and injection | 54 out of 100 |