Custom Tools

Custom Tools

Expose your own MATLAB functions as first-class MCP tools. AI agents can discover and call them directly, with full parameter validation and help text.

How It Works

1. Write your MATLAB function (.m file)
2. Describe it in custom_tools.yaml
3. The server registers it as an MCP tool at startup
4. Agents see it alongside built-in tools

Configuration

Point your config.yaml to the custom tools file:

custom_tools:
  config_file: "./custom_tools.yaml"

YAML Schema

tools:
  - name: tool_name                      # MCP tool name (what agents call)
    matlab_function: pkg.func            # MATLAB function to call
    description: "What it does"          # Shown to agents
    parameters:
      - name: param_name
        type: string                     # See Parameter Types table below
        required: true                   # or false
        description: "Parameter help"    # Optional description for agents
      - name: optional_param
        type: float
        default: 1.0                     # Default value if not provided
        description: "Optional parameter help"
    returns: "Description of return value"  # Optional

Parameter Types

YAML Type	Aliases	Python Type	MATLAB Type
string	str	str	char / string
float	number	float	double
double	—	float	double
int	integer	int	int32 / int64
bool	boolean	bool	logical
logical	—	bool	logical
list	—	list	cell
dict	—	dict	struct
any	—	Any	any

Complete Example

1. MATLAB Function (mylib/analyze_signal.m)

function result = analyze_signal(signal_path, sample_rate, window_size)
    % ANALYZE_SIGNAL  Frequency analysis of a signal file
    %
    %   result = analyze_signal(signal_path, sample_rate, window_size)
    %
    %   Returns struct with: frequencies, magnitudes, snr, peaks

    data = load(signal_path);
    signal = data.signal;

    N = length(signal);
    Y = fft(signal, window_size);
    f = (0:window_size/2-1) * sample_rate / window_size;
    mag = abs(Y(1:window_size/2)) / N;

    [peaks, locs] = findpeaks(mag, 'MinPeakHeight', max(mag)*0.1);

    result.frequencies = f;
    result.magnitudes = mag;
    result.snr = snr(signal);
    result.peaks = struct('frequencies', f(locs), 'amplitudes', peaks);
end

2. Custom Tool Definition (custom_tools.yaml)

tools:
  - name: analyze_signal
    matlab_function: mylib.analyze_signal
    description: >
      Analyze a signal file and return frequency components, SNR,
      and peak detection results.
    parameters:
      - name: signal_path
        type: string
        required: true
        description: "Path to the signal data file (.mat)"
      - name: sample_rate
        type: double
        required: true
        description: "Sample rate in Hz"
      - name: window_size
        type: int
        default: 1024
        description: "FFT window size"
    returns: "Struct with fields: frequencies, magnitudes, snr, peaks"

3. Make Sure MATLAB Can Find It

Add the directory containing your .m files to the workspace paths in config.yaml:

workspace:
  default_paths:
    - "/path/to/mylib"

4. Agent Usage

The agent now sees analyze_signal as a tool and can call it:

"Analyze the signal in data/recording.mat at 44100 Hz sample rate"

The server:

1. Validates parameters against the YAML schema
2. Calls analyze_signal('data/recording.mat', 44100, 1024) in MATLAB
3. Returns the result to the agent

Multiple Tools

tools:
  - name: analyze_signal
    matlab_function: mylib.analyze_signal
    description: "Frequency analysis of signal files"
    parameters:
      - name: signal_path
        type: string
        required: true
        description: "Path to the signal data file (.mat)"
    returns: "Frequency analysis struct"

  - name: train_model
    matlab_function: ml.train_classifier
    description: "Train a classification model"
    parameters:
      - name: dataset_path
        type: string
        required: true
        description: "Path to the training data (.mat or .csv)"
      - name: model_type
        type: string
        default: "svm"
        description: "Model type: svm, tree, knn, ensemble"
      - name: validation_split
        type: double
        default: 0.2
        description: "Fraction of data to use for validation (0-1)"
    returns: "Trained model and accuracy metrics"

  - name: process_image
    matlab_function: imgtools.enhance_image
    description: "Image enhancement pipeline"
    parameters:
      - name: image_path
        type: string
        required: true
        description: "Path to the input image"
      - name: denoise_strength
        type: double
        default: 0.5
        description: "Denoising strength (0=none, 1=maximum)"
      - name: sharpen
        type: logical
        default: false
        description: "Apply sharpening filter"
    returns: "Enhanced image saved to temp directory"

  - name: compute_statistics
    matlab_function: stats.compute_summary
    description: "Compute summary statistics for a .mat data file"
    parameters:
      - name: data_path
        type: string
        required: true
        description: "Path to the .mat file containing the data variable"
    returns: "Struct with mean, std, median, min, max, histogram data"

Loading and Registration

Custom tools are loaded at server startup:

1. The server reads the config_file path from config.yaml
2. load_custom_tools() parses the YAML and validates each tool definition using pydantic
3. For each valid CustomToolDef, make_custom_tool_handler() creates an async handler function with proper type annotations
4. FastMCP introspects the handler signature to extract parameter names and types
5. The tool is registered as an MCP tool and becomes available to agents

If the config file does not exist or tools is empty, no custom tools are registered. Invalid tool definitions log a warning and are skipped.

Parameter Descriptions

Each parameter can include an optional description field that is displayed to agents when they inspect the tool. This helps agents understand what each parameter does and what format it expects.

parameters:
  - name: signal_path
    type: string
    required: true
    description: "Path to the signal data file (.mat)"

Error Handling

If a MATLAB function throws an error during execution, the MCP server returns a structured error response to the agent, including the MATLAB error message. Ensure your functions include appropriate error checking and informative error messages.

Tips

• Function names with packages: Use pkg.func notation to call functions in MATLAB packages (e.g., +mylib/analyze_signal.m → mylib.analyze_signal)
• MEX files: Custom tools work with .mex files too — just reference the function name without the extension
• MATLAB types: Use double, logical, and other MATLAB type names in the type field for clarity; they map to their Python equivalents
• Default values: Provide sensible defaults for optional parameters so agents don't have to specify them
• Return values: Document the structure and format of what your function returns in the returns field
• Testing: Test your functions in MATLAB first before exposing them as tools