Model Configuration

LanMate ships with mainstream large models covering reasoning, multimodal, and image-understanding capabilities, and supports connecting custom model providers. Models are the foundation behind all of LanMate’s intelligent capabilities — conversations, writing, skills, and more are all driven by models.

The models you can use come from two channels: built-in models distributed centrally by your organization through the admin console — ready to use out of the box — and custom model services you connect yourself by adding them in Settings as needed. Both appear together in the model selector, and you can switch anytime.

Viewing and Switching Models

Find the model selector at the top of the conversation interface or near the input box. Click it to view all currently available models and switch the model used by the current conversation.

Models in the selector fall into two types:

Each model option shows the model name and a type tag, making it easy to tell built-in from custom models. Built-in models appear at the top of the list, custom models after them.

Setting the Default Model

If you want new conversations to automatically use a specific model, set the default model in Settings:

1. Open “Settings” and go to “Model Settings.”
2. Choose the target model from the “Default Model” dropdown.
3. It takes effect immediately — new conversations will use this model by default.

If the default model is deleted or becomes unavailable, the default automatically falls back to the first model in the list; conversations already in progress are unaffected.

Using Built-in Models

Built-in models are configured and distributed centrally by your organization via the admin console — whichever models the organization assigns to you appear on the client, ready to use as soon as LanMate is installed, with no extra configuration needed.

Notes when using:

• The features each model supports (image input, thinking mode, reasoning capability, etc.) vary by model.
• The admin console can set an enable/disable switch per model — only enabled models appear in your model selector.
• If your organization hasn’t distributed any built-in models yet, the model selector shows only the custom models you’ve added yourself.

Connecting Custom Models

If the built-in models don’t meet your needs, you can connect your own model services. Custom models are managed by provider: add a provider first, then add specific models under it, after which they’re available in the model selector.

Adding a Provider

1. Open “Settings,” go to “Model Settings,” and click Add Provider.
2. Choose an API format (OpenAI or Anthropic).
3. Fill in the Base URL (the model service’s API address) and API Key (the access credential; can be left empty for locally deployed model services).
4. After saving, the provider appears in the list.

Provider-level configuration is shared by all models under it — add it once, with no need to fill it in repeatedly for each model. The API Key is masked on screen (e.g., sk-abc…xyz) and stored encrypted locally.

Adding a Model

Add specific models under an added provider:

1. Click “Add Model” under the provider to open the form.
2. Fill in the required fields: Model Name (a display name of your choosing) and Model ID (the model ID actually called).
3. Fill in optional fields as needed: Context Length (default 128k).
4. Click “Add” to save. After the dialog closes, the new model appears in the list.

You can add multiple models under one provider, but Model IDs must not duplicate. The form can’t be submitted while required fields like Model Name and Model ID are empty.

Testing the Connection

After filling in the Base URL, it’s recommended to click the “Test Connection” button to verify the model service works:

• The button shows a loading state during the test.
• On success, it shows “Connection successful” — save and use with confidence.
• On failure, it shows “Connection failed” — check that the Base URL and API Key are correct and that the model service is running properly.

Managing Custom Models

In the model list in Settings, each model card shows the model name, Model ID, and Base URL. You can maintain them anytime:

• Edit provider: Change the name, API format, Base URL, and API Key; changes take effect for all models under it.
• Delete provider: Requires a second confirmation; deleting removes the provider together with all its models.
• Edit model: Click the edit button on the model card — the form is pre-filled with the current information; modify and save. Note that required fields can’t be emptied while editing, otherwise it can’t be saved.
• Delete model: Click the delete button on the model card and confirm. If the deleted model was the current default, the default falls back to the first model in the list; conversations in progress are unaffected, but the model is no longer selectable for new conversations.

Built-in models are maintained centrally by the admin console and can’t be edited or deleted in the client.

FAQ and Notes

Will my API Key be uploaded?

No. Model configuration is stored only in your local configuration file, and the API Key is stored encrypted locally — LanMate never uploads it to the cloud. The API Key is the key credential for accessing your third-party model account; keep it safe and don’t disclose it to others.

Where is my conversation content sent?

When using a custom model, your input is forwarded by LanMate to the corresponding third-party model service, which processes it according to its own rules. LanMate serves only as the communication link; the specific data-processing rules are governed by the agreement between you and that third party. Before connecting, it’s recommended to read the model provider’s terms of service and privacy policy first, and make sure the model source is legitimate.

How is using a custom model billed?

All costs incurred by using a custom model (token consumption, subscription fees, etc.) are paid by you directly to the third party and are unrelated to LanMate. Model calls may be triggered continuously during use — keep a close eye on your account balance and charges on the third-party platform.

Why isn’t the built-in model I need in the list?

Built-in models are configured and distributed centrally by the organization’s admin console, and only enabled models reach the client. If the model you need doesn’t appear, contact your organization’s administrator to confirm whether the model has been configured and enabled.

Usage Advice

• Prefer built-in models for a more stable service experience.
• Custom models are operated independently by third parties — for their availability, stability, and output quality, watch that third party’s service status and announcements.
• After adding a custom model, confirm the service works via “Test Connection” before putting it to use.