Waldur MCP Server
Official# Waldur MCP server
Waldur MCP server enables integration between Waldur instance and Claude Desktop by implementing the Model Context Protocol (MCP). This allows Claude to interact with your Waldur instance directly.
## Quickstart
1. Install Claude Desktop for your platform from the [Claude Desktop releases page](https://claude.ai/download)
2. Install Python 3.13 or higher.
3. Install uv package manager.
### Installation
To install the package:
```bash
pip install waldur-mcp-server
```
Or with uv:
```bash
uv pip install waldur-mcp-server
```
### Generate Waldur Token
1. Log in to your Waldur instance
2. Navigate to Username > Credentials > API Token
3. Generate a new token with appropriate token lifetime - you'll need it for configuration
### Claude Desktop MCP Server Configuration
On Windows: `%APPDATA%\Claude\claude_desktop_config.json`
On MacOS: `~/Library/Application Support/Claude/claude_desktop_config.json`
<details>
<summary>Development/Unpublished Servers Configuration</summary>
```json
{
"mcpServers": {
"waldur-mcp-server": {
"command": "uv",
"args": [
"--directory",
"C:\\waldur_service",
"run",
"waldur-mcp-server"
],
"env": {
"WALDUR_API_URL": "https://your-waldur-instance",
"WALDUR_TOKEN": "your-token"
}
}
}
}
```
</details>
<details>
<summary>Published Servers Configuration</summary>
```json
{
"mcpServers": {
"waldur-mcp-server": {
"command": "uvx",
"args": [
"waldur-mcp-server"
],
"env": {
"WALDUR_API_URL": "https://your-waldur-instance",
"WALDUR_TOKEN": "your-token"
}
}
}
}
```
</details>
### Debugging
On macOS, log files are located at `~/Library/Logs/Claude/mcp.log`
If you encounter a `spawn uvx ENOENT` error, verify that your PATH environment variable includes the Python installation directory where `uv` is installed. Alternatively, you can specify the full path to `uvx` in the configuration file, for example:
`~/Library/Frameworks/Python.framework/Versions/3.13/bin/uvx`
Since MCP servers run over stdio, debugging can be challenging. For the best debugging
experience, we strongly recommend using the [MCP Inspector](https://github.com/modelcontextprotocol/inspector).
You can launch the MCP Inspector via [`npm`](https://docs.npmjs.com/downloading-and-installing-node-js-and-npm) with this command:
```bash
npx @modelcontextprotocol/inspector uv --directory C:\\waldur_service run waldur-mcp-server
```
When debugging, you'll need to set the required environment variables. Upon launching, the Inspector will display a URL that you can access in your browser to begin debugging.
### Common issues
- Invalid token: Verify token permissions and expiration
- Connection issues: Check WALDUR_API_URL is accessible
TDQS
Scored across 7 tools
Most tools have distinct purposes: list_customers, list_invoices, list_offerings, list_projects, and list_resources each target different Waldur entities, while create_invitation and query serve unique functions. However, the five list_* tools could potentially be confused for similar listing operations, though their descriptions clarify distinct scopes.
Tool names follow a highly consistent pattern: six tools use verb_noun format (e.g., create_invitation, list_customers), and the seventh (query) aligns with a simple verb style. All names are in snake_case, making them predictable and readable across the set.
With 7 tools, the count is reasonable for a management server, covering core listing operations and key actions like invitations and queries. It feels slightly thin for comprehensive Waldur management, but each tool serves a clear purpose without bloat.
The tool set is severely incomplete for Waldur management, focusing heavily on listing operations with only one creation tool (create_invitation) and a generic query. Missing are essential CRUD operations for customers, invoices, offerings, projects, and resources (e.g., create, update, delete), leaving significant gaps that will hinder agent workflows.