agent-synapse 0.2.0

Kubernetes-like orchestration system for AI agents

Synapse

Kubernetes-like orchestration system for AI agents

Synapse is a powerful framework for orchestrating AI agents with Kubernetes-like semantics. It provides a simple YAML-based workflow definition and a rich CLI for managing agent-based applications.

Features

• 🚀 Kubernetes-inspired agent orchestration
• 📊 Built-in observability with a web dashboard
• ⚡ Fast execution with parallel agent processing
• 🔄 Dependency management between agents
• 📝 YAML-based workflow definitions
• 🔍 Tracing and logging for debugging
• 🎨 Rich CLI with beautiful output

Installation

Install from PyPI:

pip install agent-synapse

Quick Start

1. Initialize a new Synapse project:

synapse init

This creates an agents/ directory with demo agents (summarize.py and classify.py) and a workflow.yaml file.

2. Run the demo workflow:

synapse run workflow.yaml --prompt "Your text here"

3. View the dashboard (optional):

synapse serve
# Open http://localhost:8080 in your browser

4. Modify the agents in the agents/ directory to suit your needs!

Documentation

For detailed documentation, please visit the Synapse Documentation.

Key Concepts

• Agents: Independent units of work that process inputs and produce outputs
• Workflows: YAML files that define the agent graph and their dependencies
• Traces: Detailed execution logs stored in synapse_traces.db
• Dashboard: Web UI for monitoring and debugging agent executions

Examples

Check out the examples directory for sample workflows and agent implementations.

Development

1. Clone the repository:

git clone https://github.com/YakshithK/synapse.git
cd synapse

2. Set up a virtual environment:

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate
pip install -e ".[dev]"

3. Run tests:

pytest

License

This project is licensed under the MIT License - see the LICENSE file for details.

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

Support

For support, please open an issue on GitHub.