> ## Documentation Index
> Fetch the complete documentation index at: https://dshtauri.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# DSH Tauri Model

> Add model providers and configure model parameters and endpoint compatibility.

Use `dsh-tauri-model` in **Settings → Models** to manage providers, credentials, and models. It replaces the official model settings entry.

## Configure a model

1. Choose a provider from the built-in catalog or add a custom model API.
2. Enter an API key; custom APIs also need a base URL and at least one model ID.
3. Add models manually or **Fetch available models**, then review limits and capabilities.
4. Save, select the model in a session, and send a test message.

| Option | Behavior |
| - | - |
| Fetch config | Fills gaps for one model without replacing entered values |
| Configure all models | Replaces limits disclosed by the endpoint; preserves undisclosed fields |
| Image and thinking capabilities | Uses an online capability table, with cached data or naming rules as fallback; editable manually |
| Disable Developer role | Appears only for routes explicitly using `openai-completions` |
| Open config file | Opens the settings file, or its directory if the file is missing |

## Data and boundaries

* Configuration belongs to Harness settings; capability data is cached under `$DSH_HOME/dsh-tauri-model/`.
* **Open config file** uses `$DSH_HOME/settings.yaml`, or `~/.dsh/settings.yaml` when unset; it does not detect custom `settings-file` paths.
* Automatic configuration contacts provider endpoints or an online capability table and may be inaccurate.
* Do not enable the official model settings entry alongside this plugin. Keep API keys out of screenshots and issue reports.

[Model setup guide](../getting-started/configure-model) · [Source](https://github.com/dsh-tauri/deepseek-harness-desktop/tree/main/packages/dsh-tauri-model)