Skip to content

Message Types

Types for multimodal content in messages. These types enable you to pass text and images to your agents.

Overview

Zap supports multimodal content through these types:

  • TextContent: Plain text content
  • ImageContent: Image content (URL or base64-encoded data)
  • ContentPart: Union type of text or image content
  • MessageContent: Either a string or a list of content parts

Usage Example

from zap_ai import (
    Zap,
    ZapAgent,
    TextContent,
    ImageContent,
)

# Create agent with vision-capable model
agent = ZapAgent(
    name="VisionAgent",
    prompt="You are a helpful assistant that can see images.",
    model="gpt-4o",
)

zap = Zap(agents=[agent])
await zap.start()

# Send multimodal message with text and image
task = await zap.execute_task(
    agent_name="VisionAgent",
    task=[
        TextContent(type="text", text="What's in this image?"),
        ImageContent(
            type="image_url",
            image_url={"url": "https://example.com/photo.jpg"}
        ),
    ],
)

API Documentation

TextContent

zap_ai.TextContent dataclass

Text content part in a multimodal message.

Attributes:

Name Type Description
text str

The text content.
type Literal['text']

Content type identifier (always "text").

to_litellm()

Convert to LiteLLM format.

from_litellm(data) classmethod

Parse from LiteLLM format.

ImageContent

zap_ai.ImageContent dataclass

Image content part in a multimodal message.

Supports URLs (including data: URIs for base64) per LiteLLM spec.

Attributes:

Name Type Description
url str

Image URL or data: URI with base64 content.
type Literal['image_url']

Content type identifier (always "image_url").
detail ImageDetailLevel | None

Optional detail level ("auto", "low", "high") for vision models.
format str | None

Optional explicit mime type (e.g., "image/png").

to_litellm()

Convert to LiteLLM format.

from_litellm(data) classmethod

Parse from LiteLLM format.

from_url(url, detail=None) classmethod

Create from an image URL.

Parameters:

Name Type Description Default
url str

Image URL (http/https).

 required
detail ImageDetailLevel | None

Optional detail level ("auto", "low", "high") for vision models.

 None

Returns:

Type Description
ImageContent

ImageContent instance.

from_base64(data, mime_type='image/png', detail=None) classmethod

Create from base64-encoded image data.

Parameters:

Name Type Description Default
data str

Base64-encoded image data (without data: prefix).

 required
mime_type str

Image MIME type (e.g., "image/png", "image/jpeg"). Must be a valid image/* type.

 'image/png'
detail ImageDetailLevel | None

Optional detail level ("auto", "low", "high") for vision models.

 None

Returns:

Type Description
ImageContent

ImageContent instance with data: URI.

Raises:

Type Description
ValueError

If mime_type is not a valid image type.

ContentPart

zap_ai.ContentPart = TextContent | ImageContent module-attribute

MessageContent

zap_ai.MessageContent = str | list[ContentPart] module-attribute

See Also