mcp -obsidian-jsoncanvas

This MCP server implements the JSON Canvas 1.0 specification, exposing resources for working with JSON Canvas data and a set of tools to manipulate nodes, edges, and canvases. You can create, update, and delete nodes (text, file, link, group), establish edges with styling and labels, and validate or export canvases. The server also provides access to example canvases, templates, and a schema to help with validation. Use the provided tools to construct complex, infinite canvas structures, connect nodes, and validate against the official canvas schema.

Prerequisites:

• Python and the uv tool installed on your system (uv can be installed from the official source).
• Git to clone the repository or access to the package providing the MCP server.

1. Clone or download the MCP server repository for jsoncanvas.
2. Install dependencies (if any) in a virtual environment:

# Create and activate a virtual environment
python -m venv .venv
source .venv/bin/activate  # On Windows: .\.venv\Scripts\activate

# Install the package in editable mode (adjust if a setup.py/pyproject.toml exists)
pip install -e .

3. Run the MCP server using uv/uvx (as described in the repository README). For example:

uvx mcp-server-jsoncanvas

4. Verify the server starts and is listening on the expected port or endpoint as per your configuration.

Notes:

• If you need to customize output paths, set the OUTPUT_PATH environment variable.
• Ensure you have the required permissions to read/write the output directory.

Environment variables:

• OUTPUT_PATH: Directory where canvas files will be saved (default: ./output). Tweak this to point to a writable location.
• FORMAT: Default output format for canvas files (default: json). Other supported formats may include svg or png depending on export capabilities.

Common issues:

• If the server fails to start, verify that uvx is installed and that the package name (mcp-server-jsoncanvas) exists in your environment.
• Ensure you have network/file-system permissions for the OUTPUT_PATH directory.
• When using Docker or Claude Desktop integration, map the canvas data directory correctly so that outputs persist.