Vero

Vero provides an integrated analytics + AI orchestration stack with an AI agent layer powered by an MCP (Multi-Agent Coordination Protocol) server. The MCP server enables multi-agent workflows, dynamic orchestration, and performance-aware coordination, allowing agents to reason, decide, and act within your data-driven environment while remaining observable and auditable. Use the MCP layer in conjunction with the Agno AI agent interface to define agent behaviors, campaigns, and task workflows, then leverage the surrounding data warehouse, semantic models, and dashboards to monitor outcomes in real time. The provided MCP gateway coordinates tasks across agents and triggers actions across the pipeline (e.g., data ingestion, model querying, decision execution) as part of end-to-end workflows.

Prerequisites:

• Docker and Docker Compose (for the included Vero Docker setup) or a Node.js runtime if you opt to run the MCP server directly
• Git installed to clone the repository
• Access to a PostgreSQL instance for the data warehouse

Install steps:

1. Clone the repository git clone https://github.com/kpi-zone/Vero.git cd Vero

2. (Recommended) Use Docker-based deployment as documented in the repo

   • Ensure Docker is running
   • Use the provided docker-compose.yaml or docker-compose.arm64.yaml depending on your architecture
   • Copy environment example files if present and customize as needed

3. If you prefer to run the MCP server directly (Node.js based):

   • Ensure Node.js is installed (v14+ recommended)
   • Install dependencies (if a local package.json exists for the MCP server): npm install
   • Start the MCP server (example path may vary based on your repo layout): node ./ai-agent/mcp-server/index.js

4. Configure environment variables for the MCP server as needed (see mcp_config env placeholders). Ensure the PostgreSQL connection is accessible and any required API keys are provided.

5. Verify the server starts correctly and the MCP endpoint is reachable, then proceed to connect the Agno agent and the rest of the Vero stack (Cube.js, Metabase, n8n) as per the docs.

Tips and common considerations:

• If running on ARM64 (e.g., Apple Silicon), prefer the ARM64 Docker compose file and images to avoid compatibility issues with Metabase.
• Keep your ML/AI models and agent logic auditable by enabling verbose MCP logging during development (adjust MCP_LOG_LEVEL in env).
• Ensure network connectivity between the MCP server and the data warehouse (PostgreSQL) as well as any external AI services used by Agno.
• Use the provided environment variable documentation in docs/conf/environment.md to align with Vero’s configuration needs.
• When using Docker, consider setting up an nginx-proxy + Lets Encrypt workflow as shown to secure HTTPS endpoints for external access.
• Regularly pull updates from the Vero repo and review the docs/ai-agent/mcp-server.md guidance for any MCP API changes or breaking changes.