Local LLM

Local LLM

This page covers Radial Timeline's local or OpenAI-compatible AI configuration, including Ollama, LM Studio, and similar endpoints.

Local AI is a separate topic from AI Pulse Triplet Analysis. Pulse currently works best with the hosted providers and should be configured there with Anthropic, OpenAI, or Gemini for now.

What This Page Covers

Use this page when you are configuring:

• local or self-hosted AI endpoints
• OpenAI-compatible servers
• Ollama
• LM Studio
• local testing and custom model routing

AI Settings

These options appear in Settings → AI when Local/OpenAI-compatible is selected:

• Local LLM Base URL: The API endpoint. For Ollama, use http://localhost:11434/v1. For LM Studio, use http://localhost:1234/v1.
• Model ID: The exact model name your server expects.
• Custom Instructions: Additional instructions added to the start of the prompt for fine-tuning local model behavior.
• Bypass scene hover metadata yaml writes: When enabled, local analysis skips writing to the scene note and saves results in the RAW AI log instead.
• API Key (Optional): Required by some servers; usually ignored for local tools like Ollama.

Scope

Local/OpenAI-compatible configuration belongs to the AI settings layer. Treat it as infrastructure setup rather than the default guidance for Pulse or other end-user workflow pages.