Skip to content

model_registry.md

Latest commit

 

History

History
54 lines (40 loc) · 1.97 KB

model_registry.md

File metadata and controls

54 lines (40 loc) · 1.97 KB

Model Registry (Tutorial Edition)

This document lists the recommended AI setup for this tutorial stack, focusing on simplicity and cost-effectiveness.

1. Cloud Models (The "Gateway" Strategy)

Instead of managing multiple API keys (OpenAI, Anthropic, Google, etc.), we recommend using OpenRouter as a single gateway.

OpenRouter

  • Role: Universal Brain. Connects to all major models via one API key.
  • Why?: No monthly subscriptions, pay-per-token, access to top models (Claude 3.5, GPT-4o, DeepSeek V3) instantly.
  • Billing: Prepaid credits (start with $5).
  • Recommended Models:
    • Coding/Logic: anthropic/claude-3.5-sonnet (Top Tier) or deepseek/deepseek-chat (Best Value).
    • Reasoning: deepseek/deepseek-reasoner (R1).
    • General: google/gemini-2.0-flash-exp (Fast & often free on OpenRouter).

2. Local Models (Ollama)

Runs locally on your Oracle Cloud ARM64 instance. Free & Private.

Model Tag Role
Nomic Embed nomic-embed-text Essential. Powers the "Knowledge" (RAG) feature in Open WebUI.
Qwen 2.5 14B qwen2.5:14b Great balance of speed/intelligence for local chat.
Llama 3.2 llama3.2 Lightweight model for basic tasks.

📥 How to Install

Run these commands on your server to download the models:

# 1. Embeddings (Required for document search)
docker exec -it ollama ollama pull nomic-embed-text

# 2. Local Chat Model
docker exec -it ollama ollama pull qwen2.5:14b

3. Usage in Stack

  • Open WebUI:

    • Go to Settings > Connections.
    • OpenAI API: Enter https://openrouter.ai/api/v1 as the URL and your OpenRouter Key.
    • Ollama: Should be connected automatically (http://ollama:11434).

  • n8n Workflows:

    • Use the OpenAI Chat Model node.
    • Create a Credential with:
      • URL: https://openrouter.ai/api/v1
      • API Key: sk-or-...