solace-agent-mesh

Solace Agent Mesh (SAM) is a Python-based framework that enables a multi-agent AI system where specialized agents communicate over the Solace event mesh. The system handles agent orchestration, dynamic tool integration, and inter-agent collaboration, allowing you to build scalable, asynchronous workflows. With SAM, you can define agents with distinct skills, connect them to external interfaces and tools, and leverage built-in capabilities like dynamic embeds and A2A (Agent-to-Agent) communication to execute multi-step workflows. To use the server, install the solace-agent-mesh package, initialize a SAM project if needed, and run the SAM runtime via the provided CLI. The framework expects you to provide an LLM provider, tools, and agent definitions, which the ADK-based runtime will load and orchestrate over Solace.

Prerequisites:

• Python 3.10.16+ and pip
• A Solace Platform broker URL and credentials (for event mesh)
• Access to an LLM provider (e.g., OpenAI, Anthropic, or a self-hosted model) with an API key

Step-by-step installation:

1. Create and activate a Python virtual environment

python3 -m venv .venv
source .venv/bin/activate  # macOS/Linux
# Windows: .venv\Scripts\activate

2. Upgrade pip and install solace-agent-mesh

pip install --upgrade pip
pip install solace-agent-mesh

3. Verify installation

sam -v  # if the CLI is installed as part of the package, or run a basic help

4. Initialize a new project (GUI tool available)

sam init --gui

5. Run the project

sam run

6. Access the Web UI (for chat/interface) at http://localhost:8000

Notes:

• The README indicates the workflow centers around the SAM CLI (sam) and Python package installation. If your environment uses a different entry point, adapt the commands accordingly.
• Ensure your environment variables (LLM_API_KEY, SOLACE_BROKER_URL, SOLACE_BROKER_VPN) are set before running the runtime if your setup requires them.

Tips and common considerations:

• Ensure your Solace broker is reachable from the runtime environment and that the VPN/credentials are correctly configured.
• Provide your LLM provider API key via environment variables or a secure secrets store; the framework expects you to wire up the selected provider in your agent configurations.
• Use the GUI tools (sam init --gui, sam add agent --gui) to quickly scaffold agents and plugins, then customize the YAML/JSON configurations for each agent.
• If you encounter compatibility issues, verify Python version compatibility and reinstall solace-agent-mesh to ensure the proper entry points are installed.
• The project emphasizes event-driven, asynchronous execution; design agents to be decoupled and to communicate via the A2A protocol over the Solace broker.