# Help: Settings

## Configuration Sections

The Settings section allows you to customize the application according to your preferences and manage integrations with artificial intelligence services.

## 1. Application Theme

Controls the visual appearance of Evalis:

**Available options:**

- **System**: Uses the device configuration

- **Light**: Always light theme (white background, dark text)

- **Dark**: Always dark theme (dark background, light text)

## 2. Application Language

Changes the interface language:

**Supported languages:**

- **System**: Uses the language configured on the device

- **Spanish**: Interface in Spanish

- **English**: Interface in English

**Note**: Changing the language does NOT affect repository or test content.

## 3. AI Configurations

Manages integrations with artificial intelligence services.

**What are they for?**

- Create complete repositories automatically

- Generate additional questions for existing repositories

- Create quality educational content without manual writing

**Supported providers:**

- **OpenAI**: GPT-3.5, GPT-4, GPT-4o

- **Gemini**: Google models (Gemini Pro, Gemini Flash)

- **DeepSeek**: Economic models

- **Groq**: Fast inference with Llama, Mixtral

- **Custom**: Any OpenAI-compatible API

## Creating an AI Configuration

**Required fields:**

1. **Name**: Configuration identifier

2. **Provider**: Select the AI service

3. **URL**: API service address

4. **Model**: The specific model to use

5. **API Key**: Your service API key

**URLs by provider:**

- OpenAI: `https://api.openai.com/v1/chat/completions`

- Gemini: `https://generativelanguage.googleapis.com/v1beta/models/{model}:generateContent`

- DeepSeek: `https://api.deepseek.com/v1/chat/completions`

- Groq: `https://api.groq.com/openai/v1/chat/completions`

**Recommended models:**

- OpenAI: `gpt-3.5-turbo`, `gpt-4`, `gpt-4o`

- Gemini: `gemini-pro`, `gemini-1.5-flash`

- DeepSeek: `deepseek-chat`

- Groq: `llama-3.1-70b-versatile`, `mixtral-8x7b-32768`

**Advanced parameters:**

- **Max Tokens**: Maximum number of tokens in response (default: 4000)

- **Connect Timeout**: Maximum time to connect in seconds (default: 60)

- **Read Timeout**: Maximum time to receive response in seconds (default: 300)

- **Write Timeout**: Maximum time to send data in seconds (default: 60)

## Test Connection

Before using a configuration:

1. Complete all fields

2. Press "Test Connection"

3. The app sends a test request

4. Verify that "Connection successful" appears

**Common errors:**

- **Authentication error**: Incorrect API Key

- **Timeout**: Increase timeouts or verify your connection

- **Model error**: The model doesn't exist or is not available

## Getting API Keys

**OpenAI:**

1. Register at https://platform.openai.com/

2. Go to "API Keys" and create a new one

3. Add credit to your account (paid service)

**Gemini (Google):**

1. Go to https://makersuite.google.com/app/apikey

2. Create an API key

3. Free service with usage limits

**DeepSeek:**

1. Register at https://platform.deepseek.com/

2. Go to "API Keys" and create a new one

3. Economical paid service

**Groq:**

1. Register at https://console.groq.com/

2. Generate an API key

3. Free service with rate limits

## Security and Privacy

**API Key Storage:**

- Keys are securely saved in the local database

- They are NOT shared or sent to any Evalis server

- They are only used to communicate directly with the AI provider

**Recommendations:**

- Don't share your API keys

- Regenerate keys if you believe they were compromised

- Set spending limits in provider dashboards

- Review usage periodically

## Costs and Limits

**Paid services:**

- OpenAI: Pay per use (consumed tokens)

- DeepSeek: Pay per use (very economical)

**Free services:**

- Gemini: Daily request limit

- Groq: Tokens per minute limit

## Troubleshooting

**"Authentication error":**

- Verify that the API Key is correct

- Ensure the account has credit

**"Timeout" when generating:**

- Increase Read Timeout to 600 seconds

- Reduce the number of questions

- Verify your internet connection

**"Model not found":**

- Verify that the model name is exact

- Consult the provider's documentation

**"Rate limit exceeded":**

- You have exceeded the request limit

- Wait a few minutes and try again