Solace Agent Mesh is an open-source framework for building and orchestrating event-driven, multi-agent AI systems. It targets developers and enterprises seeking to create scalable, decoupled AI architectures that integrate with real-world data and systems, enabling complex, multi-step workflows through seamless agent collaboration.

How It Works

The framework leverages Solace PubSub+ Event Broker as the central event mesh for communication. It integrates Google's Agent Development Kit (ADK) for agent logic and LLM interaction with the Solace AI Connector (SAC) for runtime orchestration. Agents communicate using the A2A protocol over Solace's event mesh, facilitating task delegation, data sharing, and discovery of peer agents in a fully asynchronous, event-driven manner. This approach promotes loose coupling and high scalability for enterprise deployments.

Quick Start & Requirements

• Install via pip: pip install solace-agent-mesh
• Requires Python 3.10.16+, pip, macOS/Linux/Windows (WSL), and an LLM API key.
• Project initialization: sam init --gui within a new directory.
• Run command: sam run
• Web UI available at http://localhost:8000.
• Tutorials available for Weather Agent, SQL Database Integration, MCP Integration, and Slack Integration.

Highlighted Details

• Modular, event-driven architecture for loose coupling and scalability.
• Composable agents for complex, multi-step workflows.
• Plugin-extensible with support for custom agents and gateways.
• Built-in tools for file management, SQL queries, JQ transformations, and Plotly chart generation.
• Dynamic embeds for context-dependent information resolution.

Maintenance & Community

The project is community-driven with a list of contributors available on GitHub. Contribution guidelines are provided in CONTRIBUTING.md.

Licensing & Compatibility

Licensed under the Apache 2.0 License, permitting commercial use and integration with closed-source systems.

Limitations & Caveats

The framework relies on Solace PubSub+ Event Broker, which may require separate setup or licensing for production use. Specific LLM provider compatibility or performance tuning details are not extensively covered in the README.