Skip to content

Tutorials

Five short walkthroughs that take you from a new account to a delivered order. Do them in order the first time. The demo company “FreshPress Laundry” (from npm run db:seed) is handy for practising.

1. Set up your laundry (about 10 minutes)

  1. Choose Start free trial, fill in your company, your name, email and a password (with confirmation), and create the company. You arrive at the guided setup.
  2. Complete the company details and, if you have one, upload your logo (it appears on receipts, invoices, the tracking page and the portal).
  3. Review your branch’s opening hours, then Services: edit the sample services to match what you sell.
  4. Open Pricing: set a price for each service and garment type. Add rush tiers (for example Express) with a surcharge and faster turnaround.
  5. Open Workflows and keep or adjust the production stages. Open Operations to set rules such as tag printing, tax and receipt numbering.
  6. Open Staff and add your team. Give each person the role that fits: front desk, operators, quality control, drivers.
Not sure which role to pick? Start with the closest template. You can create custom roles later in Settings → Roles.

2. Take your first order

  1. Sign in as front desk and open New order (POS).
  2. Search for the customer, or add them (name and phone are enough).
  3. Choose a service tab and tap the garments, for example two shirts and a pair of trousers. Watch the total update.
  4. Tick Take payment now and choose Create order & take payment.
  5. You land on the order page showing Paid. Note the order number and the tracking code.

3. Tag and track the garments

  1. On the order page open the tags page. You will see one tag per garment, each with a QR code and a barcode.
  2. Print the tags, attach one to each garment, then choose Mark tags as printed.
  3. As an operator open Scan and scan a tag (or type its code, for example LD-10077-01). The garment opens.
  4. Open a garment and choose Save inspection after noting its condition. Intake cannot finish without it.
  5. In Production choose Start then Complete for each stage. The order’s status follows its garments: it shows the least advanced stage, and any rework shows as Rework.

No scanner? The Scan page also accepts typed codes, and a phone camera works once the site is on HTTPS. A USB or Bluetooth barcode scanner needs no setup: click the code box and scan.

4. Quality control and hand-over

  1. When a garment reaches quality control it appears under Quality control.
  2. Tick each checklist item that passes, leave the result on Passed and choose Record result. To fail it, untick the failing item and choose Rework required; it returns to the stage you pick.
  3. When every garment is ready the order becomes Ready. The customer’s tracking page shows it, and a notification can be sent if email or SMS is configured.
  4. At the counter open the order and choose Hand over to customer. Collect any balance first with Record payment.

5. Pickup and delivery

  1. Create an order with fulfilment set to pickup, delivery or both and enter the address, or let a customer book online at /book/your-company-code.
  2. As a delivery manager open Dispatch. Jobs needing a driver appear in the first column. Pick a driver and choose Assign.
  3. The driver opens My jobs on a phone, marks the pickup, and the item is then marked Received at branch in Dispatch. Pickup orders stay at intake until you do this.
  4. When the order is ready, assign the delivery. At the door the driver asks for the customer’s one-time code and a signature. Updates made offline are sent when the signal returns, without duplicates.

6. Read your numbers

  • The Dashboard answers “what needs attention today”: overdue orders, garments stuck for hours, low stock.
  • In Analytics compare branches and spot bottlenecks. In Reports export the figures to CSV, Excel or PDF for your accountant.
  • Optional: add an AI provider key to the server settings (ANTHROPIC_API_KEY or OPENAI_API_KEY) and use AI assistant to ask questions answered from your own data.