​
Overview

Messages are the core of LLM prompting. They define the conversation between the user and the assistant. Messages can have different roles, such as system, user, assistant, or tool. Each message role has a different meaning and purpose in the conversation.

​
Message Tags

To define a message, you can use the <message> tag, followed by a role attribute to define the role:

<message role="system">
  This is a system message.
</message>

By default, all text not wrapped into a message tag will be considered a system message, although this can be changed in the implementation.

This is a system message.

For convenience, there are tags for each specific role: <system>, <user>, <assistant>, and <tool>. These tags are equivalent to the <message> tag with the corresponding role attribute.

<system>
  You are an expert content writer with deep knowledge of {{ industry }}.
  Always write in a clear, engaging style.
</system>

<user>
  Write a blog post about {{ topic }}. Use a {{ tone }} tone.
</user>

<assistant>
  Here's a draft blog post about {{ topic }}...
</assistant>

​
Message Content

Depending on the provider, some messages can contain more than just text. For example, user messages may contain images, and assistant messages may contain tool call requests. Read more about the capabilities of your LLM provider to know what kind of content you can include in your messages.

Similar to <message> tags, you can add <content> tags to define the content of a message, followed with a type attribute to define the type of the content, which can be text, image, tool-call, or any other type supported by your LLM provider.

<user>
  <content type="text">Take a look at this image:</content>
  <content type="image">[image url]</content>
</user>

<assistant>
  <content
    type="tool-call"
    id="123"
    name="get-weather"
    arguments={{ { location: "Barcelona" } }}
  />
</assistant>

All plain text inside a message not wrapped into a content tag will automatically be considered a text content.

You can also use <content-text>, <content-image> and <tool-call> tags as shortcuts for the <content> tag with the corresponding type.

<user>
  Take a look at this image:
  <content-image>[image url]</content-image>
</user>
<assistant>
  <tool-call
    id="123"
    name="get-weather"
    arguments={{ { location: "Barcelona" } }}
  />
</assistant>

​
Image Content

Images can be included by either using <content type="image"> or <content-image>. The content of the image should contain the image as a string encoded in base64, or directly as a URL if the provider supports it.

​
Tool Call Content

Tool calls can only be included inside assistant messages, and they must contain the following attributes:

  • id: A unique identifier for the tool call.
  • name: The name of the tool to call.
  • arguments (optional): An object containing the arguments to pass to the tool.

​
Roles

​
System Messages

System messages define instructions and provide general context to the assistant.

Although the this can also be used with user messages, it is recommended to use the <system> tag to define a clear separation between the two, where system messages have a higher authority.

<system>
  You are an expert content writer with deep knowledge of {{ industry }}.
  Always write in a clear, engaging style.
</system>

​
User Messages

User messages can be used to include user input in the conversation. When including parameters straight from the user, it is recommended to use the <user> tag so that the assistant can understand the user’s input, and it can prioritize instructions accordingly.

<user>
  Write a blog post about {{ topic }}. Use a {{ tone }} tone.
</user>

​
Assistant Messages

LLMs always generate responses as assistant messages. These messages will be generated by the LLM, so you do not need to define them in your prompt, but sometimes it can be useful to fake a previous assistant response to guide the conversation. This way, when generating the response, the LLM will think it has already responded to the user, so it will follow the conversation flow.

<user>
  Hello, I am {{ name }}.
</user>

<assistant>
  Hi, {{ name }}! I am HistoryBot, your personal history assistant. What do you want to learn about today?
</assistant>

<user>
  {{ question }}
</user>

​
Tool Messages

When defining tools in the configuration, the assistant may respond with tool requests. This is usually done when the assistant needs to execute a tool to gather more information or perform actions, which should be responded with the tool’s output as a tool message. Read more about tools in your LLM’s documentation.

Like with Assistant messages, this is not usually defined in the prompt, but it can be useful to fake previous tool responses to guide the conversation.

<user>
  What's the weather like in Barcelona?
</user>

<assistant>
  <tool-call id="123" name="get-weather" arguments={{ location: "Barcelona" }} />
</assistant>

<tool id="123">
  17ºC
</tool>
ConfigurationVariables