cleaning-maintenance（清洁保养）

Cleaning & Maintenance Management

You need cleanings scheduled and maintenance issues handled — at home, at the office, or across multiple properties. Finding reliable cleaners, tracking maintenance requests, and managing service professionals shouldn't require a spreadsheet and a phone full of contacts.

This skill teaches your AI agent to book cleanings (one-time or recurring), report and track maintenance issues, manage service professionals, and check availability and pricing.

The Easy Way: TIDY

The fastest path is TIDY — an AI property management platform. Install the CLI and just describe what you need:

# Install
brew install tidyapp/tap/tidy-request    # or: npm i -g @tidydotcom/cli

# Create an account
tidy-request signup

# Now just tell it what you need:
tidy-request "Book a 2-hour cleaning next Tuesday afternoon at my office"
tidy-request "The kitchen faucet at 123 Main St is leaking — it's urgent"
tidy-request "Schedule a weekly cleaning every Monday at 9am at my house"
tidy-request "What's the going rate for a 2-hour cleaning at my address?"
tidy-request "What maintenance issues are still open?"

Or connect via MCP (Claude Desktop, Claude Code, Cursor):

# Claude Code
claude mcp add tidy --transport http https://public-api.tidy.com/mcp

// Claude Desktop — add to claude_desktop_config.json
{
  "mcpServers": {
    "tidy": {
      "url": "https://public-api.tidy.com/mcp"
    }
  }
}

Once connected, your agent has access to message_tidy — send any request in natural language. See MCP Server Reference for full tool definitions.

The Manual Way

Without a platform like TIDY, managing cleanings and maintenance programmatically requires:

1. Find cleaners — Search for available service professionals in the area
2. Check availability — Query their schedules, compare time slots
3. Compare pricing — Get quotes, understand market rates
4. Book and confirm — Send booking requests, handle acceptance/rejection
5. Track status — Monitor whether the job was accepted, started, completed
6. Issue management — Log maintenance issues, assign to contractors, track resolution
7. Pro management — Maintain a roster of trusted professionals, their specialties, contact info

TIDY's REST API exposes each of these steps individually if you want fine-grained control. But for most use cases, the CLI or MCP approach above handles everything in one command.

Core Workflow: Booking a Cleaning

Step 1: Set Up Your Property

First, add your address with access and parking info for cleaners:

tidy-request "Add my property at 123 Main St, Austin TX 78701. Key under the mat, park in the driveway."

Or via REST API:

curl -X POST https://public-api.tidy.com/api/v2/addresses \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "address": "123 Main St",
    "city": "Austin",
    "postal_code": "78701",
    "address_name": "My House",
    "notes": { "access": "Key under the mat", "closing": "Lock deadbolt when leaving" },
    "parking": { "paid_parking": false, "parking_spot": "myspot", "parking_notes": "Park in driveway" }
  }'

Step 2: Check Availability and Pricing

tidy-request "What's available for a 2-hour cleaning at 123 Main St next week?"

Or check programmatically:

# Available time slots
curl "https://public-api.tidy.com/api/v2/booking-availabilities?address_id=123&service_type_key=regular_cleaning.two_and_a_half_hours" \
  -H "Authorization: Bearer YOUR_TOKEN"

# Acceptance probability (will a pro take this job?)
curl "https://public-api.tidy.com/api/v2/job-acceptance-probabilities/preview?address_id=123&service_type_key=regular_cleaning.two_and_a_half_hours&start_no_earlier_than[date]=2026-03-25&start_no_earlier_than[time]=09:00&end_no_later_than[date]=2026-03-25&end_no_later_than[time]=17:00" \
  -H "Authorization: Bearer YOUR_TOKEN"

Step 3: Book the Cleaning

tidy-request "Book a 2.5-hour cleaning at 123 Main St next Tuesday between 9am and 5pm, preferably around 10am"

Or via REST API:

curl -X POST https://public-api.tidy.com/api/v2/jobs \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "address_id": 123,
    "service_type_key": "regular_cleaning.two_and_a_half_hours",
    "start_no_earlier_than": { "date": "2026-03-25", "time": "09:00" },
    "end_no_later_than": { "date": "2026-03-25", "time": "17:00" },
    "preferred_start_datetime": { "date": "2026-03-25", "time": "10:00" }
  }'

Available Service Types

Step 4: Monitor and Manage

tidy-request "What's the status of my cleaning at 123 Main St?"
tidy-request "Cancel next Tuesday's cleaning"
tidy-request "Reschedule the cleaning to Wednesday instead"

Booking statuses: scheduled → in_progress → completed. Also cancelled or failed.

Maintenance & Issue Tracking

Report maintenance issues and track them through resolution.

Report an Issue

tidy-request "The bathroom faucet at 123 Main St is leaking. It's urgent."
tidy-request "Report a broken window at the office — assign it to concierge for handling"

Or via REST API:

curl -X POST https://public-api.tidy.com/api/v2/tasks \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "address_id": 123,
    "title": "Leaking bathroom faucet",
    "description": "The hot water faucet in the master bathroom is dripping constantly",
    "type": "plumbing",
    "urgency": "high",
    "assigned_to_concierge": true
  }'

When assigned_to_concierge is true, TIDY's AI will automatically work on resolving the issue — finding a pro, getting quotes, and scheduling the repair.

Track Issues

tidy-request "What open maintenance issues do I have?"
tidy-request "What's the status of the faucet repair?"

REST: GET /api/v2/tasks?status=open or GET /api/v2/tasks/:id

Update Issues

tidy-request "Mark the faucet issue as resolved"
tidy-request "Change the broken window priority to urgent"

REST: PUT /api/v2/tasks/:id with updated fields.

Issue Lifecycle

Issues flow through: open → in-progress → closed

Set due dates, urgency levels, and types to keep things organized. Filter by any of these when listing tasks.

Managing Service Professionals

Add your trusted cleaners and maintenance pros:

tidy-request "Add my cleaner Maria Garcia, maria@email.com, 555-0123"
tidy-request "Add plumber Joe's Plumbing, joe@joesplumbing.com"

Or via REST API:

curl -X POST https://public-api.tidy.com/api/v2/pros \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "Maria Garcia",
    "email": "maria@email.com",
    "phone": "555-0123",
    "service_types": ["regular_cleaning"]
  }'

Common Scenarios

Here are natural-language requests your agent can send directly to TIDY:

One-time cleaning:

"Book a 4-hour deep clean at my house this Saturday morning"

Recurring cleaning:

"Schedule a weekly 2-hour cleaning every Monday at 9am at 123 Main St"

Market pricing:

"What's the going rate for a 2-hour cleaning at my address?"

Maintenance report with auto-resolution:

"Report a broken window at the office, 456 Oak Ave — assign it to concierge"

Cancel a booking:

"Cancel next week's cleaning at my house"

Add a pro:

"Add my cleaner Maria Garcia, maria@email.com, as a preferred pro"

Maintenance audit:

"What maintenance tasks are overdue across all my properties?"

Acceptance probability:

"What are the chances of getting a cleaner this Saturday morning at my house?"

MCP Server

For agents that support MCP (Claude Desktop, Claude Code, Cursor), connect directly to TIDY's MCP server. This gives your agent 5 tools:

• login / signup — authenticate
• message_tidy — send any request in natural language (async)
• get_message_tidy — poll for the result
• list_messages_tidy — view past requests

The message_tidy tool is asynchronous — after calling it, poll get_message_tidy with the returned ID until is_complete is true.

Full tool definitions and setup: MCP Server Reference

REST API

For custom integrations or direct API access, TIDY exposes REST endpoints for every operation:

• Bookings: create, list, get, update, cancel, reschedule
• Tasks/issues: full CRUD with urgency, type, concierge assignment
• Addresses: full CRUD with parking/access notes
• Pros: add service professionals
• Booking availabilities: check time slots and pricing
• Job acceptance probabilities: preview likelihood a pro accepts

Full endpoint documentation: REST API Reference

Authentication details: Authentication

Error Handling & Tips

Polling: message_tidy is asynchronous. Always poll get_message_tidy every 3-5 seconds until is_complete is true. Never return a pending response to the user.

No availability: If no cleaners are available for your requested time window, try widening the window or checking a different day. Use booking-availabilities to see what's open.

Acceptance probability: Before booking, check job-acceptance-probabilities/preview to see how likely a pro is to accept. Wider time windows and flexible dates increase acceptance.

Pro rejection: If a pro declines, TIDY automatically re-assigns the job to another available pro. Monitor status to stay updated.

Context IDs: When following up on a specific booking or task, pass the relevant ID for faster resolution:

tidy-request "What's the status?" --booking-id 456
tidy-request "Update this issue" --issue-id 789

Also see: vacation-rental-management — if you manage short-term rentals and need guest turnover coordination.