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

# Settings

> Configure application preferences and personalization options in Perplexica

Perplexica provides a settings UI where you can configure application preferences, personalization options, and manage your LLM providers.

## Accessing settings

To access the settings:

1. Click the settings icon in the Perplexica interface
2. Or navigate directly to `/settings` in your browser

<Note>
  On first launch, Perplexica displays a setup screen where you configure initial settings. You can always modify these later through the settings UI.
</Note>

## Preferences

Customize the application appearance and behavior.

### Theme

Choose between light and dark mode.

<ParamField path="theme" type="select" default="dark">
  Visual theme for the application

  **Options:**

  * Light
  * Dark

  **Default:** Dark
</ParamField>

### Measurement unit

Set your preferred measurement system for weather and other data.

<ParamField path="measureUnit" type="select" default="Metric">
  Measurement system for displaying data

  **Options:**

  * Imperial (Fahrenheit, miles, pounds)
  * Metric (Celsius, kilometers, kilograms)

  **Default:** Metric
</ParamField>

### Auto video & image search

Automatically search for relevant visual content.

<ParamField path="autoMediaSearch" type="boolean" default={true}>
  When enabled, Perplexica automatically searches for relevant images and videos based on your query

  **Default:** Enabled
</ParamField>

<Tip>
  Disabling auto media search can speed up responses if you primarily need text-based results.
</Tip>

### Show weather widget

Display weather information on the home screen.

<ParamField path="showWeatherWidget" type="boolean" default={true}>
  Shows a weather card on the home screen with current conditions

  **Default:** Enabled
</ParamField>

### Show news widget

Display recent news on the home screen.

<ParamField path="showNewsWidget" type="boolean" default={true}>
  Shows a news card on the home screen with recent headlines

  **Default:** Enabled
</ParamField>

## Personalization

Customize how the AI responds to your queries.

### System instructions

Add custom behavior or tone for the model.

<ParamField path="systemInstructions" type="textarea">
  Custom instructions that modify how the AI responds

  **Examples:**

  * "Respond in a friendly and concise tone"
  * "Use British English and format answers as bullet points"
  * "Always provide sources and explain technical terms simply"
  * "Be more formal and detailed in responses"
</ParamField>

<Accordion title="System instructions examples">
  **Concise responses:**

  ```
  Keep all responses brief and to the point. Use bullet points when listing items.
  ```

  **Academic style:**

  ```
  Respond in an academic tone with proper citations. Explain concepts thoroughly
  and include relevant background information.
  ```

  **Code-focused:**

  ```
  When discussing programming topics, always include code examples. Prefer
  Python and JavaScript examples when possible.
  ```

  **Beginner-friendly:**

  ```
  Explain technical concepts in simple terms that a beginner can understand.
  Avoid jargon and provide analogies when helpful.
  ```
</Accordion>

<Note>
  System instructions are applied to all searches and chats. They persist across sessions until you modify them.
</Note>

## Search configuration

Configure the search engine backend.

### SearXNG URL

Set the URL for your SearxNG instance.

<ParamField path="searxngURL" type="string">
  The URL of your SearxNG instance

  **Example:** `http://localhost:8080`

  **Default:** Auto-configured if using bundled SearxNG
</ParamField>

<Warning>
  If you're using the standard Perplexica Docker image with bundled SearxNG, you should not need to modify this setting.
</Warning>

See [SearxNG configuration](/configuration/searxng) for detailed setup instructions.

## Model providers

Manage your LLM providers and models.

### View configured providers

The settings UI displays all configured providers with:

* Provider name and type
* Available chat models
* Available embedding models
* Connection status

### Add a provider

<Steps>
  <Step title="Click Add Provider">
    In the settings UI, click the "Add Provider" button
  </Step>

  <Step title="Select provider type">
    Choose from:

    * Ollama
    * OpenAI
    * Anthropic
    * Gemini
    * Groq
    * LM Studio
    * Lemonade
    * Transformers
  </Step>

  <Step title="Configure settings">
    Enter the required configuration:

    * API keys (for cloud providers)
    * Base URLs (for local/custom endpoints)
    * Provider name (to identify multiple instances)
  </Step>

  <Step title="Test connection">
    Perplexica automatically tests the connection and fetches available models
  </Step>

  <Step title="Save">
    Save the provider configuration
  </Step>
</Steps>

### Edit a provider

1. Find the provider in the settings list
2. Click "Edit"
3. Modify the configuration
4. Save changes

<Note>
  Editing a provider re-tests the connection and refreshes the available models list.
</Note>

### Remove a provider

1. Find the provider in the settings list
2. Click "Delete" or "Remove"
3. Confirm deletion

<Warning>
  Removing a provider makes its models unavailable. Searches configured to use models from this provider will fail.
</Warning>

### Add custom models

For providers that don't auto-detect all models:

1. Select the provider
2. Click "Add Custom Model"
3. Enter:
   * **Model Name**: Display name
   * **Model Key**: API identifier
   * **Type**: Chat or Embedding
4. Save the model

See [Models and providers](/configuration/models-and-providers) for provider-specific configuration details.

## Search modes

While not configurable in settings, Perplexica offers three search modes you can select per query:

<Accordion title="Speed mode">
  **Best for:** Quick answers and simple questions

  * Fastest response time
  * Uses lighter models
  * Suitable for straightforward queries
</Accordion>

<Accordion title="Balanced mode">
  **Best for:** Everyday searches

  * Good balance of speed and quality
  * Recommended for most queries
  * Default mode
</Accordion>

<Accordion title="Quality mode">
  **Best for:** Deep research and complex topics

  * Most thorough responses
  * Uses more powerful models
  * Takes longer but provides detailed answers
</Accordion>

## Focus modes

Perplexica supports different focus modes for specialized searches:

* **All** - General web search
* **Academic** - Search academic papers and scholarly sources
* **Writing** - Optimized for content creation and writing assistance
* **Wolfram Alpha** - Mathematical and computational queries
* **YouTube** - Video search
* **Reddit** - Search Reddit discussions

<Note>
  Focus modes are selected per search and are not configured in settings.
</Note>

## Data and privacy

### Search history

Perplexica stores your search history locally:

* All searches are saved in your browser's local storage
* History persists across sessions
* You can view past searches in the Library
* No search data is sent to external servers (except for API calls to your configured LLM providers)

<Tip>
  Your searches are completely private when using local LLM providers like Ollama.
</Tip>

### File uploads

Uploaded files are stored:

* In the Docker volume mounted at `/home/perplexica/uploads`
* Files persist across container restarts
* Supported formats: PDF, text files, images

## Configuration persistence

All settings are stored in:

* **Location:** `/home/perplexica/data/config.json`
* **Persistence:** Stored in Docker volume `perplexica-data`
* **Backup:** Automatically saved when modified

<Warning>
  If you remove the Docker volume, all settings will be lost and you'll need to reconfigure Perplexica.
</Warning>

## Environment variables override

Settings configured via environment variables:

* Take precedence over UI settings on startup
* Are automatically applied when the container starts
* Create provider entries in the configuration
* Can be viewed but not edited through the UI (edit the environment variables instead)

See [Environment variables](/configuration/environment-variables) for details.

## Exporting and importing configuration

To backup or migrate your configuration:

### Export configuration

```bash theme={null}
# Copy from Docker volume
docker cp perplexica:/home/perplexica/data/config.json ./config-backup.json
```

### Import configuration

```bash theme={null}
# Copy to Docker volume
docker cp ./config-backup.json perplexica:/home/perplexica/data/config.json
# Restart container
docker restart perplexica
```

## Next steps

* [Configure LLM providers](/configuration/models-and-providers)
* [Set up SearxNG](/configuration/searxng)
* [Environment variables reference](/configuration/environment-variables)