Trackings MCP Server
# Trackings MCP Server
MCP server for Claude Desktop integration with trackings.ai.
## Installation
```bash
uvx --from git+https://github.com/peterw/trackings-mcp trackings-mcp
```
## Configuration
Add to your Claude Desktop config (`~/Library/Application Support/Claude/claude_desktop_config.json`):
```json
{
"mcpServers": {
"trackings": {
"command": "uvx",
"args": ["--from", "git+https://github.com/peterw/trackings-mcp", "trackings-mcp"],
"env": {
"TRACKINGS_API_KEY": "your-api-key",
"TRACKINGS_API_URL": "https://your-convex-deployment.convex.site"
}
}
}
}
```
## Available Tools
- `list_projects` - List all projects
- `list_scans` - List scans for a project
- `list_scan_runs` - List scan runs for a project or scan
- `get_scan_run` - Get details for a scan run
- `get_scan_results` - Get consolidated keyword results for a run
- `get_credits` - Get current credit balance
- `create_scan` - Create a scan configuration
- `trigger_scan_run` - Trigger a scan run
TDQS
Scored across 8 tools
Every tool has a clearly distinct purpose targeting specific resources and actions in the tracking/scanning domain. For example, create_scan configures scans, trigger_scan_run executes them, get_scan_results retrieves keyword data, and get_scan_run fetches run details—no overlap or ambiguity exists.
All tools follow a consistent verb_noun pattern with snake_case, such as create_scan, get_credits, list_projects, and trigger_scan_run. This uniformity makes the toolset predictable and easy for agents to navigate without confusion.
With 8 tools, the server is well-scoped for managing scans, projects, and runs in a tracking system. Each tool earns its place by covering distinct operations like creation, listing, triggering, and retrieval, avoiding bloat or thin coverage.
The toolset provides strong CRUD/lifecycle coverage for scans and projects, including create, list, get, and trigger operations. A minor gap exists in update/delete functionality for scans or projects, but agents can work around this for core workflows.