portfolio-manager-go

The Portfolio Manager MCP server provides an API-driven interface to value and analyze a personal investment portfolio across equities, forex, commodities, cash, bonds, and cryptocurrencies, with the added capability of leveraging LLM-based insights. It stores data locally in a LevelDB-backed datastore and supports calculating portfolio statistics such as IRR, market value, and price paid, while also offering OpenAPI-compliant endpoints for easy integration with other systems. The server integrates with Gemini for analytics-derived insights and can generate trading ideas, correlation matrices, and benchmarking against user-defined benchmarks. Users can interact with the server through the built-in CLI, REST/OpenAPI endpoints, and a UI frontend when paired with the web UI in development workflows.

Prerequisites:

• Go installed (as the project is Go-based) and Make utility
• Git to clone repositories
• Optional: Node.js for the UI if you plan to run the web frontend locally

Installation steps:

1. Clone the repository: git clone https://github.com/rodionlim/portfolio-manager-go.git cd portfolio-manager-go

2. Ensure you have Go 1.24.3 or higher installed as required by the project.

3. Build and install the application using Make: make

4. Run the built binary to start the server. If you built locally via make, the binary is typically named portfolio-manager. Start it with: ./portfolio-manager -config custom-config.yaml

5. If you prefer to run via Docker (as indicated in the MCP config):

   • Pull and run the image: docker run -i rodionlim/portfolio-manager-go
   • Pass in configuration via a config file or environment as needed by your setup.

6. Optional: Start the UI (if you want the web frontend): cd web/ui npm install npm run dev

7. Validate installation by querying the API or CLI as described in the README (e.g., list positions, backup/restore commands, etc.).

Environment variables:

• GEMINI_API_KEY: API key for Google Gemini AI service. If set, this will override geminiApiKey in the config.
• ANALYTICS_SCHEDULE: Cron expression to schedule SGX report collection. If set, overrides analytics.schedule in config.

Proxmox VE helper scripts are provided for easy deployment in LXC containers. See the README for how to install and configure environment variables in systemd unit files. If using Google Drive backup/restore features, OAuth2 desktop credentials are recommended for personal accounts. The database can be wiped with make clean-db for development workflows. The server exposes CLI commands for position management and database backup/restore, with offline local access required for backup/restore operations.