# AI Configuration

### Simple talker  preset

The setup has predefined parameters that you can alter.

1. Choose the LLM of your preference. Don't feel any pressure, you may change it anytime later or even use different LLMs for different parts of the scenario (configured in chaingraph).
2. Review and edit system prompt. This is the main instruction for your agent. By default, it is configured to act as an AI assistant. Tailor it your needs. 

{% hint style="info" %}
In the prompt try to give the agent clear instructions on what its purpose and role are. Describe their personality, tone of voice, spice it up with catchphrases. Experiment  with response rules, reasoning and models.
{% endhint %}

<figure><img src="/files/GtOyNVKPw3bK0YyzEROc" alt="" width="563"><figcaption></figcaption></figure>

3. The last four parameters refer to chat history

`MESSAGES_COUNT` defines the number of messages stored in the agent's chat history. Adjusting this parameter balances context depth with memory efficiency.

`TOKENS_LIMIT` sets the maximum token count the model can process in one interaction. A higher limit allows for longer and more detailed exchanges but requires more computational resources.

`TEMPERATURE_CONTROL` controls the randomness of the AI's responses. Lower values (e.g., 0.3) produce more precise and deterministic outputs, while higher values (closer to 1) allow for creative and varied answers.

`MAX_TOKENS` specifies the maximum token output for each response.

### Advanced mode

You should understand that all of these parameters correspond to nodes in chaingraph. For example, the pre-defined set of parameters corresponds to Agent Profile node. You do not have to use it, though. Just unlink the node and use configure your own graph in the way you like.

<figure><img src="/files/6ipfmvhY5dQvDLkULB7z" alt="" width="169"><figcaption></figcaption></figure>

AI configuration tab makes your life easier when dealing with tangled graphs and tens of parameters. If you want to set them up in the section, add the keys in the node called "Template Param" and you'll be able to configure those keys in AI Configuration tab.&#x20;

<div><figure><img src="/files/YINEwkHT3Gp3q2pLCUAg" alt=""><figcaption></figcaption></figure> <figure><img src="/files/BFs9QfbmCXlQRkvVAufq" alt=""><figcaption></figcaption></figure></div>

Don't forget to Save the agent every time you make changes to AI configuration.

<figure><img src="/files/vU8hVMlAAH7fMzTaZ6sO" alt="" width="563"><figcaption></figcaption></figure>


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://archai.gitbook.io/archai/create-your-own-ai/ai-configuration.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.