mcp-server-jobber
v1.0.1
Published
MCP Server for Jobber API - Read clients, jobs, quotes, invoices via natural language
Downloads
250
Maintainers
Readme
mcp-server-jobber
An MCP server that connects AI assistants to your Jobber account. Ask questions about your clients, jobs, quotes, and invoices in plain English — no clicking through dashboards.
Built for home service pros: plumbers, landscapers, cleaners, electricians, HVAC techs, and anyone running their business on Jobber.
What can you do with it?
Once connected, just ask your AI assistant things like:
- "Which invoices are past due?" — Instantly see every unpaid invoice with client name, amount, and due date.
- "Show me all active jobs for this week" — Get a clear list of what's scheduled, who the client is, and where the work is.
- "Give me the full history for John Smith" — Pull up a client's contact info, properties, jobs, quotes, and invoices in one shot.
No more switching tabs. No more scrolling. Just ask.
Available tools
| Tool | Description |
|------|-------------|
| list_clients | Search and list your clients by name, email, or phone number |
| list_jobs | List jobs filtered by status — active, completed, or upcoming |
| list_quotes | List quotes filtered by status — draft, awaiting response, approved, converted, or archived |
| list_invoices | List invoices filtered by status — draft, awaiting payment, paid, or past due |
| get_client_details | Get everything about a client: contact info, properties, custom fields, recent jobs, quotes, and invoices |
| get_business_summary | Quick snapshot of your business — total clients, jobs, quotes, and invoices |
Prerequisites
- Node.js 18 or later
- A Jobber developer app with Client ID and Client Secret (create one here)
- Set
http://localhost:3456/callbackas the Redirect URI in the Jobber Developer Center
Installation
Step 1 — Authorize your Jobber account (one time)
JOBBER_CLIENT_ID=your_client_id \
JOBBER_CLIENT_SECRET=your_client_secret \
npx mcp-server-jobber-authThis opens your browser to log in to Jobber and grant access. Tokens are saved locally at ~/.mcp-jobber/tokens.json and refresh automatically.
Step 2 — Connect to your AI assistant
Claude Desktop
Add this to your claude_desktop_config.json:
{
"mcpServers": {
"jobber": {
"command": "npx",
"args": ["-y", "mcp-server-jobber"],
"env": {
"JOBBER_CLIENT_ID": "your_client_id",
"JOBBER_CLIENT_SECRET": "your_client_secret"
}
}
}
}Claude Code
Run this in your terminal:
claude mcp add jobber -- npx -y mcp-server-jobberThen set the environment variables JOBBER_CLIENT_ID and JOBBER_CLIENT_SECRET in your Claude Code configuration.
How it works
- Read-only — This server only reads data from Jobber. It never creates, updates, or deletes anything.
- OAuth2 — Secure authentication with automatic token refresh and rotation.
- GraphQL — Uses the Jobber GraphQL API with cursor-based pagination and built-in rate limit handling.
- stdio transport — Standard MCP transport, compatible with any MCP client.
Security
- Your Jobber credentials are stored locally on your machine (
~/.mcp-jobber/tokens.json). - Tokens refresh automatically. If a refresh fails, you simply re-run the auth command.
- No data is sent anywhere except directly to the Jobber API.
License
MIT
