Build Agents, Not Infrastructure

Agentuity Python SDK

Visit https://agentuity.com to get started with Agentuity.

The Agentuity Python SDK is a powerful toolkit for building, deploying, and managing AI agents in Python environments. This SDK provides developers with a comprehensive set of tools to create intelligent, event-driven agents that can process various types of content, communicate with each other, and integrate with external systems.

Key Features

• Multi-Agent Architecture: Build and orchestrate multiple interconnected agents that can communicate and collaborate.
• Event-Driven Design: Respond to various triggers including webhooks, cron jobs, SMS, voice, email, and more.
• Rich Content Handling: Process and generate multiple content types including JSON, text, markdown, HTML, and binary formats (images, audio, PDFs).
• Persistent Storage: Built-in key-value and vector storage capabilities for maintaining state and performing semantic searches.
• Observability: Integrated OpenTelemetry support for comprehensive logging, metrics, and tracing.
• Cross-Runtime Support: Works seamlessly with both Node.js and Bun runtimes.

Use Cases

• Building conversational AI systems
• Creating automated workflows with multiple specialized agents
• Developing content processing and generation pipelines
• Implementing intelligent data processing systems
• Building AI-powered APIs and services

Getting Started

To use this SDK in a real project, you should install the Agentuity CLI.

Mac OS

brew tap agentuity/tap && brew install agentuity

Linux or Windows

See the Agentuity CLI repository for installation instructions and releases.

Once installed, you can create a new project with the following command:

agentuity new

Development Setup

Prerequisites

• Python (3.11 or higher)
• uv (latest version recommended)

Installation

Clone the repository and install dependencies:

# Clone the repository
git clone https://github.com/agenuity/sdk-py.git
cd sdk-py

# Install dependencies using uv (recommended)
uv install

Local Development

Create a minimal agentuity.yaml file in the root of the project:

project_id: proj_mv0w67c4851eX0oTglYLzP3OgS2BjN54
name: test
agents:
  - id: agent_HaDpiH67c4851eISzbAWfZqwLtnpguW6
    name: myfirstagent

Create a test agent:

mkdir -p agents/myfirstagent
touch agents/myfirstagent/agent.py

Create a test agent:

def run(request, response, context):
    return response.text("Hello, world")

Run the test server:

uv run test.py

Hit the test endpoint:

curl -v http://localhost:3500/agent_HaDpiH67c4851eISzbAWfZqwLtnpguW6 --json '{"trigger":"manual","contentType":"text/plain","payload":"aGkK"}'

You can also use the run endpoint which emulates the production one:

curl -v http://localhost:3500/run/agent_HaDpiH67c4851eISzbAWfZqwLtnpguW6 --json '{"hello":"world"}'

License

See the LICENSE file for details.