About Dart MCP Server
Dart MCP Server exposes task, document, space, and dartboard operations from the Dart project management platform to AI assistants. It calls the Dart SDK through a Python subprocess and communicates with MCP clients over stdio.
Key features
- 01Create and update tasks
- 02Look up dartboards and default statuses
- 03List and create folders
- 04Create documents and spaces
- 05Delete spaces
Use cases
- 01Create or modify Dart tasks
- 02Look up space and dartboard details
- 03Create project documents or reports
- 04Manage workspaces and folders
MCP support
As an MCP server, it provides clients with tools for Dart tasks, documents, spaces, and dartboards over stdio.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
Tools
Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.
Details
From the project READMEdart-mcp-server
MCP server that bridges the Dart project management platform to AI assistants. The server is TypeScript but delegates all Dart API calls to a Python subprocess using the dart-sdk package.
What It Does
Exposes Dart's task management, document, space, and dartboard operations as MCP tools. Each tool call spawns a Python process that imports dart-sdk, executes the operation, and returns JSON results over stdout. Communicates with MCP clients over stdio.
Status
| Area | State |
|---|---|
| MCP transport | stdio |
| Language | TypeScript (server) + Python (Dart SDK calls) |
| MCP SDK | @modelcontextprotocol/sdk ^1.5.0 |
| Dart SDK | dart-sdk (pip, called via subprocess) |
| Node.js | >= 16.0.0 |
| Python | >= 3.8.0 |
| License | MIT |
MCP Tools
| Tool | What It Does |
|---|---|
create_task | Create a task with title, description, priority, status, assignees, tags |
update_task | Update an existing task's status, title, or description by DUID |
get_default_status | Return default status DUIDs for a dartboard |
get_default_space | Return the default space DUID |
get_dartboards | List available dartboards |
get_folders | List folders in a space |
create_folder | Create a new folder |
create_doc | Create a document or report |
create_space | Create a new workspace |
delete_space | Delete a workspace by DUID |
Setup
Via Smithery
npx -y @smithery/cli install @jmanhype/dart-mcp-server --client claude
Manual
git clone https://github.com/jmanhype/dart-mcp-server.git
cd dart-mcp-server
npm install # also runs postinstall: creates .venv and pip installs dart-sdk
npm run build
npm start
Environment Variables
| Variable | Required | Description |
|---|---|---|
DART_TOKEN | Yes | Dart API token (get from your Dart profile) |
PYTHONPATH | No | Path to Dart SDK if not in default location |
PYTHON_PATH | No | Path to Python executable (defaults to system Python) |
Verifying the Python Environment
python -c "import dart; print(dart.__version__)"
If this fails, the MCP server will not be able to execute any tools.
Architecture
src/
index.ts — MCP server, tool handlers, Python subprocess management
test.ts — Basic test runner
Each tool handler builds a Python script as a string, spawns python3 -c "<script>", and parses the JSON output. The Python process has a 120-second timeout.
Limitations
- Every tool call spawns a new Python process (no persistent connection to Dart API)
- The
.envpath is hardcoded to/Users/speed/dart-tools/.envin the source - No input validation beyond what the Dart SDK provides
- No pagination support for list operations
- Error messages from the Python subprocess are passed through as-is
- The
postinstallscript creates a virtualenv, which may fail in some CI environments
Dependencies
| Package | Version | Purpose |
|---|---|---|
@modelcontextprotocol/sdk | ^1.5.0 | MCP server protocol |
dotenv | ^16.4.7 | Environment variable loading |
typescript | ^5.3.3 | Build toolchain |
dart-sdk | (pip) | Python client for Dart API |
License
MIT
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.