# Request

The **Request** tile is a universal way to work with AI in Tila. You can enter any text, choose a model, specify the desired response format, and get results in seconds. Suitable for a wide range of scenarios — from generating text and code to processing files and creating visual content.

## What does it do?

Lets you enter any request in natural language (question, command, instruction) and automatically determines the appropriate response type: text, image, list, document, etc.

## When to use?

Use this tile when you want to quickly formulate a task in free form, without complex settings.

## Main tile elements

### Request field

Enter a question or instruction. You can use free text or precise commands.

### Response format

Choose the desired result format if needed (image, video, document, etc.). You can let the AI decide.

### Model selection

Set the model manually (for example, GPT-4o, Claude 3.5 Sonnet, Gemini 1.5 Pro) or leave **AUTO** so Tila will choose the best one. More: [Choosing a language model](/content/help/models/choose/llm.html).

### Run button

Click the blue arrow button or press Enter to start generation.

## How to use

1. Create a **Request** tile via “+” on the left panel.
2. Formulate your task.
3. Attach a file if needed.
4. Specify the response format or leave it up to AI.
5. Choose the model or leave AUTO.
6. Click the button to start generation.

## Example use

> “Draw a cute bear for a birthday tile”

The AI will determine that image generation is required and return a picture.

## Deleting a tile

- **Windows/Linux** — press Delete
- **macOS** — press Backspace

## Recommendations

- Use specific instructions: “Create”, “Compare”, “Generate”, “Write as an expert…”. The more detailed the task, the more predictable the result.
- Combine with other tiles for more complex scenarios.
- If there are errors, try regenerating, rephrasing your request, or changing the model.