A truck arrives with 50 new laptops. They have to be in the CMDB by Monday. You're not going to open the New Computer form 50 times. GLPI ships three serious mechanisms for bulk onboarding — asset templates, CSV import, and GLPI Agent discovery — and a fourth, mass edit, for fixing what the first three got wrong. Used in the right order, 50 laptops takes an hour. Used badly, it takes a week.
Start with an asset template
An asset template in GLPI is a pre-filled blank. Open Assets > Computers > Templates (the template selector in the top-right corner of the list view), or Setup > Templates for full management. Create a template named after the deployment batch — for example 2026-Q2-Latitude-laptops — and pre-fill everything that's the same across the batch:
- Manufacturer, Model, Type, OS family, OS version
- Default location (warehouse, before deployment) and status (In stock)
- Supplier and contract reference
- Default group (technical owner: IT-hardware)
- Warranty period and supplier-warranty start date if known
Set the auto-naming pattern on the template — for example LT-2026-#### — and GLPI will generate sequential names. Whatever import method you choose next will inherit everything from this template, leaving you to supply only the per-device differences: serial number, MAC, and assignee.
CSV import for known data
When the supplier sends an Excel spreadsheet with serial numbers and MACs, that's the path of least resistance. Install the Data Injection plugin (the standard import path in modern GLPI). The plugin lets you upload a CSV, map columns to GLPI fields, save the mapping as a "model" for reuse, and apply the asset template you created above so missing fields get default values.
Three CSV gotchas that bite first-timers:
- Encoding. Excel-saved CSV defaults to Windows-1250 on Slovak locales. GLPI expects UTF-8. Save as CSV UTF-8 from Excel, or convert with
iconvbefore import. - Delimiter. Slovak Excel writes semicolons, GLPI defaults to commas. Set the delimiter explicitly in the import dialog.
- Existing-record detection. The plugin can update existing records or skip them — set it explicitly. If you set it to "skip" and a serial already exists from an agent scan, your CSV row is silently dropped.
GLPI Agent for discovery
For assets that will boot before you import — typical when you image laptops in batches — install the GLPI Agent during the image build. The first time the device contacts the GLPI inventory endpoint, it self-registers with everything it knows about itself: hostname, serial, MAC, CPU, RAM, OS, installed software. Far more accurate than a CSV you typed by hand.
Pair with the CSV approach: import a CSV with assignment data (who gets which laptop) first, then let the agent fill in the technical details when each machine boots. Agent data merges into existing records by serial number, so your CSV-defined assignee survives.
Mass edit for the inevitable corrections
After import, you'll discover something wrong consistently across the batch — wrong default location, wrong group, status that should have been Deployed not In stock. Don't open 50 forms again. Filter the asset list by the template name or import batch, select all, and use the bulk action menu: Modify. You can change any field on the selected assets in one operation.
Mass edit also handles assignment to people: select the laptops, choose Add a user, pick — but only do this if assignment was 1:1 to a single person, otherwise revert to the CSV approach with per-row user references.
Validate before announcing the batch is done
The post-import check that separates a usable CMDB from a half-empty one: filter Assets > Computers by Inventory number = empty, Serial number = empty, User = empty, and Location = warehouse default. Anything that shows up is unfinished. Fix it before the next batch arrives, because each unfinished asset becomes invisible work that surfaces during the next audit — by which time nobody remembers what it was.