Ollama + Cloudflare + Obsidian BMO Chatbot

Course: Software Engineering — Final Project

A self-hosted AI server running Ollama, tunneled through Cloudflare, and connected to Obsidian via the BMO Chatbot plugin.

---

How It Works

Obsidian BMO Plugin → Cloudflare Tunnel → Ollama (localhost:11434)

Your prompts go from Obsidian through a secure Cloudflare tunnel to Ollama running on your local machine. No cloud AI provider needed.

---

Dependencies

Tool	Purpose
Ollama	Runs the local AI model
cloudflared	Exposes Ollama over a public HTTPS URL
Obsidian	Note-taking app
BMO Chatbot Plugin	Connects Obsidian to the Ollama API

---

Setup

1. Start Ollama

ollama pull llama3
ollama serve

2. Create a Cloudflare Tunnel

cloudflared tunnel login
cloudflared tunnel create ollama-tunnel

Edit ~/.cloudflared/config.yml:

tunnel: <YOUR_TUNNEL_ID>
credentials-file: /home/<USER>/.cloudflared/<YOUR_TUNNEL_ID>.json

ingress:
  - hostname: ollama.yourdomain.com
    service: http://localhost:11434
  - service: http_status:404

cloudflared tunnel route dns ollama-tunnel ollama.yourdomain.com
cloudflared tunnel run ollama-tunnel

3. Configure BMO Plugin

1. Obsidian → Settings → Community Plugins → install BMO Chatbot
2. Set REST API URL to https://ollama.yourdomain.com
3. Set Model to llama3
4. Leave API Key blank

---

Project Structure

/
├── README.md
├── requirements.txt
├── config/
│   └── cloudflared-config.yml
├── docs/
│   ├── architecture.md
│   └── dependencies.md
└── src/

---

Submission Checklist

• README.md
• requirements.txt
• docs/ — architecture & dependencies
• src.tar.gz — source files (tar -czf src.tar.gz src/)