automateyournetwork / IOS_XE_RAPTOR

An AI assistant for Cisco IOS XE that uses RAPTOR Tree Multi-Query RAG and Ollama so you can 'chat' with your Cisco IOS XE device

IOS_XE_RAPTOR

An AI assistant for Cisco IOS XE that uses RAPTOR Tree Multi-Query RAG and Ollama so you can 'chat' with your Cisco IOS XE device

Getting started

Clone the repo

Update the testbed.yaml file locally to reflect your device; it currently uses the DevNet Always On Sandbox

Bring up the server

docker-compose up

Visit Ollama and download your model(s)

http://localhost:3002

Gear / settings button

Models

Download phi, mistral, llama2, gemma, etc

Start Packet Raptor

http://localhost:8503

Usage

This has been tested with a variety of IOS XE configurations and works best with larger data sets. For smaller devices / configurations I would recommend Packet Buddy instead.

The tool will download the Instructor-XL model dynamically, be patient, the first time you launch it, in order to provide free open source embeddings