> For the complete documentation index, see [llms.txt](https://railnology.gitbook.io/railnology-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://railnology.gitbook.io/railnology-docs/raillie-ai-assistant/what-is-raillie.md).

# What Is Raillie?

Raillie is Railnology's AI compliance assistant. Its mission is to turn rail knowledge into confidence — giving your team fast, accurate answers to compliance and operational questions without requiring a regulatory expert on call at all times.

Raillie is powered by a retrieval-augmented generation (RAG) pipeline that combines a curated regulatory knowledge base with the Anthropic Claude AI model. When you ask Raillie a question, it searches the relevant knowledge sources, retrieves the most applicable content, and generates a response grounded in that material.

### What Raillie Knows

Raillie's global knowledge base covers:

* FRA regulations — 49 CFR Parts 213, 218, 228, 240, 242, and more
* Transport Canada rail operating rules and standards
* ARTF regulatory framework (Mexico)
* FRA advisory circulars and interpretive guidance
* Operating rules frameworks including GCOR and NORAC structures

If your organization is on the Professional plan or above and has uploaded documents to its Private Knowledge Base, Raillie can also search those documents.

### What Raillie Is Not

Raillie is a compliance aid — not a substitute for qualified legal or regulatory counsel. Every Raillie response includes the following disclaimer:

*"Raillie is a compliance aid — verify all answers against official regulatory sources before acting on them."*

Always verify Raillie's answers against the official CFR text or your railroad's approved procedures before making operational decisions.