冰朔 8739805f99
Some checks failed
Auto-update PR branches / Update open PR branches (push) Has been cancelled
CI / Frontend Static Quality Checks (push) Has been cancelled
CI / Frontend Tests & Coverage (push) Has been cancelled
CI / Rust Tests & Quality Checks (push) Has been cancelled
CI / Linux build verification (push) Has been cancelled
Release (Alpha) / Compute alpha version (push) Has been cancelled
Release (Alpha) / Build release artifacts (push) Has been cancelled
Release (Alpha) / GitHub Release (alpha) (push) Has been cancelled
Release (Alpha) / Update docs and release pages (push) Has been cancelled
Deploy docs / Build VitePress site (push) Has been cancelled
Deploy docs / Deploy to GitHub Pages (push) Has been cancelled
光湖开源源码快照 · Tolaria AGPL 分叉基线 · 独立更新链
2026-07-05 17:45:16 +08:00

1.0 KiB

Configure AI Models

Source: guides/configure-ai-models.md URL: /guides/configure-ai-models

Configure AI Models

Use model providers when you want chat over note context without giving an agent vault-write tools.

Local Models

Local model targets are for tools such as Ollama and LM Studio. They usually need a base URL and model ID, and they usually do not need an API key.

API Models

API model targets are for hosted providers such as OpenAI, Anthropic, Gemini, OpenRouter, or another OpenAI-compatible endpoint.

Tolaria does not store provider API keys in vault settings. Choose one of the supported key paths:

  • Save the key locally on this device.
  • Read the key from an environment variable.
  • Use no key for local providers that do not require one.

Test The Connection

After adding a provider, use the test action in Settings. A successful test means Tolaria reached the endpoint and the model replied.

Select The Target

Once configured, choose the model from the AI target selector or set it as the default AI target in Settings.