jobkeepr-mcp
Summary: Manage a field service business end to end — read jobs, customers, schedules, estimates, invoices, payments and line items, and create/reschedule/cancel/close jobs, customers, events, estimates, catalog items and line items.
Business context: Get business name, timezone, working hours, price catalog, lead sources, and resolved date ranges for today/tomorrow/this week/last month/this year.
Read jobs & schedule: Query jobs by date/status/customer/type/source; view scheduled jobs, calendar events, and open availability slots.
Read customers: Search customers by name/phone/email, optionally with job counts and total paid.
Read financials: Retrieve estimates, invoices (with unpaid balances), payments, and job line items for revenue/margin analysis.
Reference data: List job types, lead sources, and catalog items with pricing.
Create jobs: Book jobs for existing or new customers, with notes or billable line items, service type, address, duration, and lead source.
Create customers & events: Add customers standalone or link calendar events (time off, meetings, walkthroughs) to customers/crew.
Create estimates & catalog items: Build quotes with catalog or custom pricing; add services/products to the price catalog.
Add line items: Attach billable services/products to an existing job.
Modify jobs: Reschedule date/time/duration, cancel scheduled jobs, and mark jobs complete (cancellation and completion are destructive-flagged).
Set lead source: Record or change how a customer found the business.
Limitation: Cannot text customers — SMS is deliberately excluded.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@jobkeepr-mcp What's on my schedule tomorrow?"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Jobkeepr MCP server
Connect Jobkeepr to Claude, ChatGPT, or any other MCP client, so an AI assistant can read and manage your field service business: jobs, customers, scheduling, estimates, invoices and payments.
Jobkeepr is field service management software for small contractor crews — handymen, TV installers, plumbers, electricians, HVAC techs — running one to ten people.
"What's on my schedule tomorrow?"
"Which service made me the most money last month?"
"Book Sarah Chen for a TV mount Thursday at 2pm."
"Mark today's jobs complete."Two ways to connect
Remote (recommended). If your client supports remote MCP servers, point it straight at the hosted endpoint and skip this package entirely:
https://jobkeepr.com/mcpIt supports OAuth, so you sign in with your normal Jobkeepr login and approve access in the browser. Nothing to install.
stdio (this package). For clients that only speak stdio, this bridges to the same server using an API key.
This package is not on npm. Install it straight from GitHub using the
github:dandyer/jobkeepr-mcpspecifier shown below. Plainnpx jobkeepr-mcpwill not resolve. npm compiles the TypeScript for you on install, so there is nothing extra to run.
Related MCP server: Jobber MCP Server
Setup
Create an API key in Jobkeepr under Settings → API Keys. It's shown once, so copy it then.
Claude Desktop
Add to claude_desktop_config.json:
{
"mcpServers": {
"jobkeepr": {
"command": "npx",
"args": ["-y", "github:dandyer/jobkeepr-mcp"],
"env": {
"JOBKEEPR_API_KEY": "jk_live_your_key_here"
}
}
}
}Claude Code
claude mcp add jobkeepr --env JOBKEEPR_API_KEY=jk_live_your_key_here -- npx -y github:dandyer/jobkeepr-mcpOr connect to the remote endpoint directly:
claude mcp add --transport http jobkeepr https://jobkeepr.com/mcpTools
Reading
Tool | What it does |
| Business name, timezone, working hours, and today's date with resolved ranges for "this week", "last month" and so on |
| Jobs filtered by date, status, customer, type or source |
| Search customers, optionally with job counts and totals |
| Scheduled jobs, calendar events, and open time slots |
| Estimates and their status |
| Invoices and unpaid balances |
| Payments received |
| Line items across jobs, for margin and mix analysis |
| Job types, lead sources, and the price catalog |
Writing
Tool | What it does |
| Book a job for a customer |
| Add a calendar event that is not a job, like time off or a walkthrough |
| Add a customer |
| Create an estimate |
| Add a service or product to the price catalog |
| Add billable line items to a job |
| Record how a customer found the business |
| Move a job to a new date or time |
| Cancel a scheduled job |
| Mark jobs complete |
The financial tools require the VIEW_FINANCIALS permission, and each write tool requires its matching permission. Tools you cannot use are not listed to you.
What it will not do
No tool here can text your customers. Sending SMS is deliberately excluded: it is the one action that reaches a real person and cannot be undone. Ask your assistant to draft a message and send it yourself from Jobkeepr.
Writes take effect immediately. There is no staging step and no undo, which is why cancel_job and close_jobs are marked destructive so your client prompts before running them. close_jobs can act on many jobs at once, so check the set before approving it.
Configuration
Variable | Required | Default | Notes |
| yes | — | From Settings → API Keys. Starts with |
| no |
| Override only for self-hosted or staging. |
An API key acts on one account and carries owner-level access. Revoke it any time in Settings → API Keys. To give an assistant narrower, per-user access instead, use the remote endpoint with OAuth.
How it works
This package is a proxy, not a reimplementation. The tools live on the hosted server; this bridges stdio to it. That means the tool set here is always current and this package rarely needs a release.
Development
npm install
npm run build
JOBKEEPR_API_KEY=jk_live_... node dist/index.jsTest interactively with the MCP Inspector:
JOBKEEPR_API_KEY=jk_live_... npx @modelcontextprotocol/inspector node dist/index.jsLinks
MIT licensed.
Available Tools
19 toolsadd_line_itemsAdd line items to a jobAInspect
Add billable line items to an EXISTING job. Use when the user wants to add services/products to a job that's already scheduled (e.g., 'add a $50 cable management charge to the Smith job'). For brand-new jobs, use create_job with lineItems instead. Pass unitPrice when the user states a price; omit it to fall back to catalog matching by name.
Inputs: lineItems (required), jobId, customerName.
cancel_jobCancel a jobADestructiveInspect
Cancel a scheduled job by setting its status to CANCELED. No data is deleted. Provide customerName (finds most recent scheduled job) OR jobId (exact match). Use customerName when user mentions a customer.
Inputs: jobId, customerName.
close_jobsMark jobs completeADestructiveInspect
Mark jobs as complete/done. ALWAYS use this tool when user wants to: close jobs, complete jobs, mark jobs done, finish jobs. Examples: 'close today's jobs', 'mark the Smith job done', 'complete yesterday's jobs'. Do NOT use get_jobs first - this tool finds and closes jobs in one step.
Inputs: date, jobIds, customerName.
create_catalog_itemAdd a price catalog itemAInspect
Add a new service or product to the catalog with pricing. Use when user mentions pricing for services/products. This is the ONLY place to store pricing information.
Inputs: name (required), price (required), cost, itemType, +2 more.
create_customerCreate a customerAInspect
Add a new customer without scheduling a job. For new customers WITH a job, use create_job instead — it handles creating the customer inline. Extract ALL contact details the user provides (name, phone, email, address).
Inputs: name (required), email, phone, source, +2 more.
create_estimateCreate an estimateAInspect
Create an estimate/quote for a customer. When the user specifies a flat rate or custom price (e.g. 'total $250', 'charge $300 flat'), pass lineItems with that custom unitPrice — do NOT use catalog pricing. Only fall back to catalog pricing when no price is specified. For flat rates, use a single line item with the service name and the stated price.
Inputs: customerName (required), email, notes, phone, +3 more.
create_eventAdd a calendar eventAInspect
Add a calendar event that is NOT a job: time off, a walkthrough, a meeting, a personal appointment, a block of time. Use this when the user says 'event', 'block off', 'time off', 'meeting', or describes something on the calendar that is not billable work. For scheduling billable work for a customer, use create_job instead. An event can optionally be linked to an existing customer and/or a crew member. Pull the date and time from anywhere in the message and resolve relative dates using the DATE REFERENCE.
Inputs: date (required), title (required), time, notes, +7 more.
create_jobCreate a jobAInspect
Schedule/create a new job for a customer. Pull the date and time from anywhere in the user's message, including pasted or forwarded text (e.g. 'scheduled for Wednesday the 24th at 9am'). Resolve relative dates like 'next monday' or 'the 24th' using the DATE REFERENCE in your instructions. Only ask for a date/time when none appears anywhere in the message — never re-ask for one the user already provided. If the customer is new (doesn't exist yet), pass their contact details here — do NOT call create_customer separately. This tool handles creating the customer AND the job in one step. When user says 'create customer X and schedule a job', use ONLY this tool.
Inputs: date (required), time (required), customerName (required), email, +8 more.
get_business_contextGet business context and date referenceARead-onlyIdempotentInspect
Get this business's name, timezone, working hours, price catalog, lead sources, and — most importantly — today's date plus the resolved YYYY-MM-DD ranges for "today", "tomorrow", "this week", "last week", "this month", "last month", and "this year" in the business's own timezone and week-start convention. Call this FIRST when a request involves any relative date, and pass the returned values to the dateFrom/dateTo arguments of the other tools rather than computing dates yourself. Also call it BEFORE creating a job, estimate, or line items: the returned catalog supplies the catalogItemId and price for each line, and jobSources lists the only accepted values for create_job's required source.
Inputs: no named parameters.
get_customersSearch customersARead-onlyIdempotentInspect
Search or list customers. Use for customer lookups, finding contact info, or customer-related questions.
Inputs: limit, query, includeStats.
get_estimatesList estimatesARead-onlyIdempotentInspect
Fetch estimates/quotes. Use for pending proposals, conversion tracking, estimate follow-ups.
Inputs: limit, dateTo, status, dateFrom, +1 more.
get_invoicesList invoicesBRead-onlyIdempotentInspect
Fetch invoice records. Use for billing questions, unpaid balances, invoice status tracking.
Inputs: dateFrom (required), jobId, limit, dateTo, +2 more.
get_jobsSearch jobsARead-onlyIdempotentInspect
Fetch job records with filters. Use for any question about jobs, work performed, job types, job sources, or schedules. Returns raw job data - you calculate totals, counts, and comparisons.
Inputs: dateFrom (required), limit, dateTo, status, +3 more.
get_line_itemsList job line itemsARead-onlyIdempotentInspect
Fetch line items from jobs (services/products sold). Use for questions about which items make money, most popular services, item profitability, catalog usage.
Inputs: dateFrom (required), limit, dateTo, jobStatus, +1 more.
get_paymentsList paymentsARead-onlyIdempotentInspect
Fetch payment records. Use for revenue questions, income tracking, payment history. Returns raw payment data - you sum and analyze.
Inputs: dateFrom (required), jobId, limit, dateTo, +1 more.
get_reference_dataList job types, sources, and price catalogARead-onlyIdempotentInspect
Fetch lookup tables: job types, job sources, catalog items. Use when you need to know what types/sources exist, catalog pricing, or to map IDs to names.
Inputs: types (required), search, includeArchived.
get_scheduleGet schedule and availabilityARead-onlyIdempotentInspect
Get scheduled jobs, calendar events (time off, meetings, walkthroughs), and optionally available time slots. Use for 'what's on the calendar', 'when am I free', scheduling questions.
Inputs: dateTo (required), dateFrom (required), includeAvailability.
reschedule_jobReschedule a jobAInspect
Change the date, time, or length of an existing job. Provide customerName (finds most recent scheduled job) OR jobId (exact match). Supply at least one of newDate, newTime, or durationMinutes. When moving the job, provide both newDate and newTime where possible. When the user only wants the appointment to block off more or less time (e.g. 'make Jon's appointment 3 hours', 'that job needs 90 minutes'), pass durationMinutes alone and leave the date and time out so the job stays where it is.
Inputs: jobId, newDate, newTime, customerName, +1 more.
set_customer_sourceSet a customer's lead sourceAInspect
Set or change how a customer found the business (their lead source). The source is stored on the customer, not on individual jobs, and every job they book counts toward it.
Inputs: sourceName (required), customerName (required).
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 19 tools
Most tools have clearly distinct purposes (create_job vs create_event vs create_customer are explicitly differentiated, and add_line_items vs create_job+lineItems is clarified). Minor overlap exists between get_jobs and get_schedule and between get_business_context and get_reference_data, which both return catalog/lookup data, but the descriptions give workable guidance.
Every tool follows a strict snake_case verb_noun convention (get_jobs, create_job, reschedule_job, close_jobs, cancel_job, add_line_items, set_customer_source). No camelCase or naming outliers, making the set highly predictable.
19 tools is on the heavier side but reasonable for a multi-entity domain (jobs, customers, estimates, invoices, payments, catalog, scheduling). Each tool maps to a distinct operation, though a few could plausibly be consolidated (e.g. get_reference_data vs get_business_context).
Job lifecycle is well covered (create, read, reschedule, close, cancel) and estimates/catalog have creation paths, but notable gaps exist: no update_customer, no update/delete for estimates, invoices, or line items, and no way to record payments or convert estimates to jobs. Read-only surfaces for invoices/payments leave dead ends.
Maintenance
Related MCP Connectors
Run a home-service business from your AI: jobs, customers, invoices, estimates, texts.
Run a service business from your AI agent: CRM, leads, scheduling, invoicing, jobs and more.
Job management for service companies: look up, book, tag, assign and note jobs from your AI.
Read FieldPulse customers, jobs, estimates, invoices, assets and projects; create and update jobs.
Related MCP Servers
- -licenseNot gradedqualityNot gradedmaintenanceEnables AI agents to interact with ServiceFusion's field service management platform through standardized MCP operations. Supports complete customer and job management including creating work orders, managing customer records, and accessing service data with OAuth 2.0 authentication.-
- AlicenseAqualityFmaintenanceEnables AI assistants to access and manage Jobber field-service data including clients, jobs, invoices, and quotes through natural language interactions.679 npmMIT
- AlicenseAqualityBmaintenanceMCP server for the Housecall Pro API, letting AI assistants read and write Housecall Pro data—customers, jobs, invoices, estimates, scheduling, and more—through natural language.4174 npmMIT
- AlicenseAqualityBmaintenanceEnables AI assistants to read field-service business data through a read-only MCP integration, including jobs, invoices, quotes, schedules, clients, and draft client messages. It handles authentication, token refresh, throttling, schema versioning, and account-currency formatting for live Jobber accounts.7MIT