> For the complete documentation index, see [llms.txt](https://docs.obside.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.obside.com/obside-chat/automate-your-trading/deploy-an-ai-trader.md).

# Deploy an AI trader

## Overview

An **AI trader** is an **autonomous** agent powered by a large language model (LLM). Unlike a bot that follows strict rules, an AI trader **interprets the market context** (price, volume, news, indicators…) and **makes its own decisions** within the boundaries you set.

{% hint style="info" %}
You create an AI trader the same way as any other agent — by **describing it in** [**Obside Chat**](/obside-chat/introduction.md). There's no dedicated button: just spell out the mandate, and the agent's **AI** (and usually **News**) capability is enabled automatically. See [Agent capabilities](/agents/capabilities.md).
{% endhint %}

You define:

* **The assets** it can trade.
* **The risk management** (position size, stop-loss, maximum exposure).
* **The style** (intraday, swing, momentum, macro, contrarian…).
* **The sources** it can consult (price, indicators, news, X/Twitter).

It then decides **when to enter, when to exit, and how much to risk**, based on its read of the market.

{% hint style="info" %}
An AI trader runs in **paper trading** (demo) mode by default. It's an excellent way to evaluate its behavior before any live deployment.
{% endhint %}

## How to phrase your request

Define a **clear mandate**: asset universe, risk constraints, expected style. The more precise the framing, the more consistent the AI's decisions will be with your intent.

## Example requests

{% hint style="info" %}
**Example #1 — Risk-conscious agent**

"Deploy an AI trader that paper trades MNQ and MGC with strict risk management. On each trade, set a stop-loss and take-profit, and size the position so that hitting the stop costs at most 0.25% of capital. Only enter when the reward/risk ratio is clearly favorable. Use only market orders."
{% endhint %}

{% hint style="info" %}
**Example #2 — Momentum hunter**

"Deploy an AI trader that paper trades TSLA, NVDA, PLTR, MSFT, COIN and SMCI, focusing on stocks showing strong intraday momentum and unusual volume. Risk at most 0.5% of capital per trade. Use tight stops and trail profits with the move. Exit any position by end of day."
{% endhint %}

{% hint style="info" %}
**Example #3 — Macro agent**

"Deploy an AI trader that paper trades MES and MGC, based on macro catalysts and news. Watch economic releases, central bank announcements and major geopolitical events. Take positions when a clear directional thesis emerges. Risk 0.5% of capital per trade and hold positions 1 to 5 days."
{% endhint %}

***

## AI trader vs trading bot: when to choose what?

|                     | **Trading bot**             | **AI trader**                                 |
| ------------------- | --------------------------- | --------------------------------------------- |
| **Logic**           | Strict, deterministic rules | Contextual decisions                          |
| **Reproducibility** | Identical on every run      | Variable depending on context                 |
| **Backtest**        | Relevant and reliable       | Limited — the AI exploits the present context |
| **Best for**        | Systematic strategies       | Reading the market and adapting               |

{% hint style="success" %}
The two approaches are complementary. You can absolutely have several agents running in parallel, each with its own mandate.
{% endhint %}