Help & Examples
Complete guide to using Local (Offline) & Online (Cloud) AI models, plus ticket management commands.
← Back to Chat
🤖 AI Models: How to Use Local vs Online
🏠 Local • Offline • Gemma-2B

Local / Offline Model

Runs 100% locally inside the container via llama.cpp CPU inference. Zero external API keys required, private, fast, and operates completely offline.

Force Local Prefix: @local <query>
  • @local how to debug memory leaks in Python
  • @local explain deadlock vs race condition
  • @local what causes segmentation faults in C++?
  • @local show all open tickets
☁️ Cloud • Online • Gemini

Online / Cloud Model

Powered by Google Gemini with live tool execution, Google Search, StackOverflow issue cross-referencing, multi-modal screenshot analysis, and deep reasoning.

Force Cloud Prefix: @cloud <query>
  • @cloud investigate CVE-2024-3094 vulnerability
  • @cloud analyze this stack trace and suggest root causes
  • @cloud find solutions for AWS RDS pool exhaustion
  • Attach error screenshots using the 📷 camera button
⚡

Smart Automatic Routing (Default)

If you don't prefix with @local or @cloud, Tandion automatically routes simple questions, greetings, and ticket queries to the fast Local model, and complex bug troubleshooting to Cloud Gemini. If external cloud services are ever unavailable, the system automatically falls back to your local model!

📋 Ticket Management Commands

Create Ticket

  • Create a ticket: login fails after update
  • Report a bug: dashboard loads forever
  • Open a ticket for VPN not connecting

Update Status

  • Close ticket 12
  • Resolve ticket #7
  • Mark ticket 5 as in progress

Update Priority

  • Set ticket 9 to critical
  • Change priority of ticket 3 to P1
  • Make ticket 14 low priority

View & Search

  • Show all open tickets
  • List critical tickets
  • Search tickets for "login error"