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

# Choose how the assistant responds

> Set the tone, level of exploration, instructions, and language for future answers.

Open your profile menu, then **Settings → Avatar**. The answer settings here guide how the assistant communicates. They are separate from your profile photo, the app’s appearance, and your health history.

## Style

Move the **Style** slider between **Precise**, **Balanced**, and **Creative**. It requests less or more variation in wording and ideas. **Balanced** is the default; use **Reset Style** to return to it.

A more creative setting is not permission to invent health facts. A precise setting is not a guarantee that an answer will be correct or identical every time. The effect can vary with the model and request.

## Form of answer

Choose the approach that helps you understand an answer:

| Option           | What it asks the assistant to do                                                                 |
| ---------------- | ------------------------------------------------------------------------------------------------ |
| Extremely strict | Stay close to established facts and explicit evidence, state uncertainty, and avoid speculation. |
| Default          | Give a clear, balanced answer with detail suited to your request.                                |
| Epistemic        | Explain the evidence, assumptions, uncertainties, and competing explanations.                    |
| Curious          | Explore useful alternatives and clarifying questions without inventing facts.                    |

These are communication preferences, not different levels of medical authority.

## Instructions

Use **Instructions** for guidance you want the assistant to remember when composing future answers. The field accepts up to **5,000 characters**.

For example: “Start with a short summary. Explain unfamiliar terms. Use metric units unless I ask otherwise.”

Edit or clear the field when your preference changes. These instructions do not rewrite answers already in your history, and they do not override safety requirements.

## Active rules

Use **Active rules** for short, separate preferences. Each rule can contain up to **500 characters**.

* Type into **New rule** and press **Enter** to add it. Moving focus outside the new-rule row also saves a non-empty draft.
* Use **Shift + Enter** for a line break while drafting. **Escape** clears the new draft.
* Edit an existing rule in its field. Use the arrows to move it up or down, or the delete control to remove it.

Rules are presented to the assistant in the order shown. Keep them clear and consistent; ordering is not a guarantee that conflicting instructions will be resolved exactly as you expect. Like other preferences, rules cannot override safety requirements.

## Preferred response language

Choose a language to ask the assistant to answer in it. **Automatic** follows the language of each conversation. Use the reset control beside a selected language to return to Automatic.

You can still ask for a different language in a particular message. Important units, medication names, citations, and quotations may stay in their original form when translating them would change their meaning.

This setting controls written answers. It does not change the app’s interface language or choose a spoken voice. For voice and read-aloud controls, see [Voice](/voice).

## Try a simple request

After changing a preference, send a new message and check whether the result helps. You can ask for a shorter explanation or a different format directly in the conversation without changing every setting.

If the answer is wrong, correct the information rather than trying to fix it only by moving the Style slider.