Local LLM setup (Ollama / LM Studio)
For maximum privacy, run a model on your machine and select it in VaultCharts. The assistant lives on its own AI chat page (similar to ChatGPT) — it is not tied to “opening a chart first.”
Steps
- Install Ollama or LM Studio and pull a model that fits your RAM.
- In VaultCharts, open Models (assistant / AI settings).
- Pick a local provider from the dropdown (Ollama, LM Studio, or the local option your build lists).
- Pick a model ID from the list, or add a model ID if yours is not listed yet.
- Open the AI chat page and ask a grounded question (e.g. scan patterns / summarize a setup). The assistant can use tools against data already in your vault.
- Verify any levels or plans yourself; confirm writes when prompted.
If connection fails, use Test connection in Models settings and see AI chat FAQ.
When to use local vs cloud
Prefer local when you want privacy / no API bill.
Prefer cloud when you want stronger reasoning and less hardware.
Trade plans and chart data stay on your device either way.