You can run and interact with the ADM1 MCP Server to perform natural-language assisted anaerobic digestion modeling using the ADM1 standard. It connects an MCP client to a Python-based simulation engine, enabling feedstock interpretation, reactor configuration, and professional reports through conversational prompts and tools.

How to use

Interact with the ADM1 MCP Server through your MCP client to model anaerobic digestion with ADM1. Begin by configuring the client to run the local Python-based server, then perform a sequence of steps to describe your feedstock, set reactor parameters, run the simulation, and analyze results. You can leverage a suite of core simulation tools, analysis tools, and utility tools to generate detailed stream properties, inhibition analyses, biomass yields, and professional reports.

How to install

# 1. Install dependencies
git clone https://github.com/puran-water/adm1-mcp.git
cd adm1-mcp
python -m venv venv
venv\Scripts\activate  # Windows
# source venv/bin/activate  # macOS/Linux
pip install -r requirements.txt

# 2. Configure environment
cp .env.example .env
# Edit .env and add your Google API key:
# GOOGLE_API_KEY=your_google_api_key_here

# 3. Configure Claude Desktop (example)

{
  "mcpServers": {
    "adm1-mcp": {
      "command": "C:\\path\\to\\your\\venv\\Scripts\\python.exe",
      "args": ["C:\\path\\to\\adm1-mcp\\server.py"],
      "env": {
        "MCP_TIMEOUT": "600000"
      }
    }
  }
}

Configuration File Locations:

• Windows: %APPDATA%\\Claude\\claude_desktop_config.json
• macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
• Linux: ~/.config/claude/claude_desktop_config.json

4. Restart Claude Desktop

After updating the configuration, restart Claude Desktop to load the ADM1 server.

## Configuration and usage notes

The ADM1 MCP Server integrates with Claude Desktop or other MCP clients. You typically run the server locally and connect via the client using a stdio configuration that points to your Python interpreter and the server script. You can adjust timeouts to accommodate longer simulations and set environment variables as needed for your environment.

## Troubleshooting and tips

Common issues include environment activation problems, API key misconfigurations, and connection problems between the MCP client and the local server. Ensure the virtual environment is activated, dependencies are installed, and the client configuration accurately references the Python executable and server script. If simulations have convergence issues, try a different integration method, adjust the time step, and verify feedstock charge balance before running simulations.

## Notes on usage flow

The typical flow involves preparing feedstock descriptions, optionally generating kinetic parameters, validating charge balance, configuring reactor properties, running one or more simulations (for up to three reactor configurations), and then examining stream properties, inhibition analyses, biomass yields, and reports. You can reuse the same server instance for multiple simulations by altering parameters between runs.

## Professional reporting and outputs

Results are presented in publication-quality formats with no scientific notation artifacts, including detailed visualizations and exportable reports that cover process health, KPI dashboards, and technical methodology. Data is formatted for clear presentation and downstream analysis.

## Available tools

### describe\_feedstock

Generate ADM1 state variables from a natural language description of the feedstock.

### describe\_kinetics

Generate both state variables AND kinetic parameters from a natural language description of the feedstock.

### set\_flow\_parameters

Configure the influent flow rate and basic simulation timing parameters (flow\_rate, simulation\_time, time\_step).

### set\_reactor\_parameters

Set reactor-specific parameters such as temperature, HRT, and integration method for a given reactor index (1-3).

### run\_simulation\_tool

Execute the ADM1 simulation using the currently configured parameters.

### get\_stream\_properties

Analyze detailed properties of specified streams (influent, effluent, biogas).

### get\_inhibition\_analysis

Provide process health analysis with inhibition factors and optimization recommendations.

### get\_biomass\_yields

Calculate biomass yields and key performance metrics for the simulation.

### validate\_feedstock\_charge\_balance

Verify thermodynamic consistency of the defined feedstock.

### check\_nutrient\_balance

Analyze C:N:P ratios for process optimization.

### get\_parameter

Retrieve current parameter values from the simulation state.

### set\_parameter

Modify specific simulation parameters; note that this invalidates prior results.

### generate\_report

Create professional simulation reports with plots and tables.

### reset\_simulation

Reset all simulation parameters to their default values.