Getting started¶

The directory where you clone this repository is referred to as the working directory or WD hereinafter.

Install¶

Install uv and just. To install the project with its minimal (runtime) dependencies in a virtual environment, run the following in the WD.

just install

To also install the dependencies required for developing and testing PriorisMCP, use the install-all target instead.

just install-all

Standalone usage¶

PriorisMCP can be started standalone as an MCP server with the stdio transport by running the following in the WD.

uv run prioris-mcp

Alternatively, it can be started using the streamable-http or http transports by setting the PRIORIS_MCP_TRANSPORT environment variable — see Configuration.

Connect it to Claude¶

PriorisMCP exposes a stdio-transport entry point (prioris-mcp), which is what both Claude Desktop and Claude Code expect for a locally-run MCP server. There are two ways to launch it, depending on whether you are running from a published release or a local clone.

Using uvx (published package, no clone required)¶

Once PriorisMCP is published to PyPI, uvx can fetch and run it on demand without installing anything into a project environment first.

Claude Code

claude mcp add prioris-mcp -- uvx prioris-mcp

Claude Desktop — add the following to your claude_desktop_config.json (macOS: ~/Library/Application Support/Claude/claude_desktop_config.json; Windows: %APPDATA%\Claude\claude_desktop_config.json):

{
    "mcpServers": {
        "prioris-mcp": {
            "command": "uvx",
            "args": ["prioris-mcp"]
        }
    }
}

Using uv (local clone, e.g. for development)¶

If you are working from a local clone of this repository (see Install above), point uv run at the WD with --directory so it resolves regardless of Claude's own working directory.

Claude Code

claude mcp add prioris-mcp -- uv run --directory /absolute/path/to/prioris-mcp prioris-mcp

Claude Desktop

{
    "mcpServers": {
        "prioris-mcp": {
            "command": "uv",
            "args": ["run", "--directory", "/absolute/path/to/prioris-mcp", "prioris-mcp"]
        }
    }
}

Instead of uv run --directory ... prioris-mcp, you may also point command directly at the virtual environment's script, e.g. /absolute/path/to/prioris-mcp/.venv/bin/prioris-mcp.

To pass configuration (see Configuration) to the server, add an env block alongside command/args in either config, e.g. "env": {"PRIORIS_MCP_LOG_LEVEL": "DEBUG"}, or pass --env KEY=VALUE flags to claude mcp add.

Test with the MCP Inspector¶

The MCP Inspector is an official Model Context Protocol tool that can be used by developers to test and debug MCP servers. This is the most comprehensive way to explore the MCP server.

To use it, you must have Node.js installed. The best way to install and manage node as well as packages such as the MCP Inspector is to use the Node Version Manager (or, nvm). Once you have nvm installed, you can install and use the latest Long Term Release version of node by executing the following.

nvm install --lts
nvm use --lts

Following that, run the MCP Inspector and PriorisMCP by executing the following in the WD.

npx @modelcontextprotocol/inspector uv run prioris-mcp

This will create a local URL at port 6274 with an authentication token, which you can copy and browse to on your browser. Once on the MCP Inspector UI, press Connect to connect to the MCP server. Thereafter, you can explore the tools available on the server.

You can, alternatively, launch the MCP Inspector by running just launch-inspector.

Testing and coverage¶

To run the provided set of tests with a coverage report, execute the following in the WD.

just test-coverage

This will generate something like the following output.

Name    Stmts   Miss    Cover   Missing
---------------------------------------
TOTAL     98      0  100.00%

See the Contributing guide for the full development workflow.