ag-ui-protocol

Python SDK for the Agent-User Interaction (AG-UI) Protocol.

ag-ui-protocol provides Python developers with strongly-typed data structures and event encoding for building AG-UI compatible agent servers. Built on Pydantic for robust validation and automatic camelCase serialization for seamless frontend integration.

Installation

pip install ag-ui-protocol
poetry add ag-ui-protocol
pipenv install ag-ui-protocol

Features

🐍 Python-native – Idiomatic Python APIs with full type hints and validation
📋 Pydantic models – Runtime validation and automatic JSON serialization
🔄 Streaming events – 16 core event types for real-time agent communication
⚡ High performance – Efficient event encoding for Server-Sent Events

Quick example

from ag_ui.core import TextMessageContentEvent, EventType
from ag_ui.encoder import EventEncoder

# Create a streaming text event
event = TextMessageContentEvent(
    type=EventType.TEXT_MESSAGE_CONTENT,
    message_id="msg_123",
    delta="Hello from Python!"
)

# Encode for HTTP streaming
encoder = EventEncoder()
sse_data = encoder.encode(event)
# Output: data: {"type":"TEXT_MESSAGE_CONTENT","messageId":"msg_123","delta":"Hello from Python!"}\n\n

Multimodal user message

from ag_ui.core import UserMessage, TextInputContent, ImageInputPart, InputContentUrlSource

message = UserMessage(
    id="user-123",
    content=[
        TextInputContent(text="Please describe this image"),
        ImageInputPart(
            source=InputContentUrlSource(
                value="https://example.com/cat.png",
                mime_type="image/png",
            )
        ),
    ],
)

payload = message.model_dump(by_alias=True)
# {"id": "user-123", "role": "user", "content": [...]}

BinaryInputContent is deprecated. Use modality-specific input parts (ImageInputPart, AudioInputPart, VideoInputPart, DocumentInputPart) with InputContentDataSource or InputContentUrlSource.