Importing data

Import files: CSV, vendor PDFs and shelf photos

What you can upload, how to gather it all in one batch, and how the column matching works.

The guided import reads what you already have rather than making you retype it. This article covers the first half: getting your files in and confirming how their columns were understood. For what happens after that, see Review and finish an import. If what you already have is a spreadsheet, what you gain by moving off one is the case for doing this at all.

Uploading files and matching columns do not create inventory items. Items are added only when you approve rows in the review.

Who can import

PI and Researcher roles. Junior Researchers do not see the import option.

Where to start

  1. On the Dashboard, the "Set up your lab" card has a Start setup button.
  2. Or in the top bar, open Add Inventory and choose Import inventory.
  3. Or go to Tools in the sidebar and open Import Inventory.

What you can upload

TypeNotes
CSV spreadsheetsExport from Excel or Google Sheets as CSV.
PDF documentsVendor order confirmations, packing slips.
PhotosJPG, JPEG, PNG, HEIC, or HEIF. A shelf, a freezer box, a set of labels.

Excel files are not supported. Save the sheet as CSV first.

Upload everything at once

Click Choose files and add the whole batch: the spreadsheet, the PDFs, the photos, together. They are read into a single review rather than one import per file.

This is worth doing deliberately. Labs usually have their inventory scattered across several sources, and importing them one at a time means reconciling the overlaps yourself afterwards. Bringing them in together lets the review handle it in one pass.

You can also add another file later and have it fold into the same session.

The templates

If you would rather start from a template than adapt your own sheet, use Download a template on the upload screen. There is one for each kind of item, and each has every column Labsistant reads for that kind:

  • Reagent: Item name, Kind, Quantity, Unit, Alert at, Quantity type, Hazard level, CAS number, Vendor, Catalog number, Category, Lot number, Received date, Expiration date, Lot cost, Sterile, RNase-free, Storage location, Box, Position in box, Ownership, Owner email, Owner group, Notes.
  • Supply: Item name, Kind, Quantity, Unit, Alert at, Quantity type, Vendor, Category, Sterile, RNase-free, Storage location, Box, Position in box, Ownership, Owner email, Owner group, Notes.
  • Sample: Item name, Quantity, Unit, Quantity type, Sample type, Collection date, Sample source, Biohazard, Already thawed, Storage location, Box, Position in box, Ownership, Owner email, Owner group.

Whatever file you upload, put one item per row, with a header row first. Next to the templates, the upload screen says what a row needs:

  • Needed now for a reagent: Item name, Quantity, Unit, Lot number (or No lot yet) and Storage location. A supply needs the same, without the lot.
  • Can be added later: Hazard level, CAS number, Catalog number, Vendor, Received date, Expiration date, Lot cost, Alert at and Notes.

Leave a column blank when you do not have the value. A few columns accept only certain values:

  • Hazard level is one of none, low, moderate, high or extreme, or leave it blank and add it later.
  • Dates are written year-month-day, such as 2026-01-15. If your dates are written another way, the import asks how to read them.
  • Quantity type is discrete for things you count and continuous for things you measure. Leave it blank and Labsistant works it out from a unit it knows.
  • Ownership is shared, group or personal.
  • Sterile, RNase-free, Biohazard and Already thawed are yes or no.
  • Notes is free text, written to the lot's notes.
  • Kind is read row by row in a reagent or supply file: reagent or chemical makes the row a reagent, supply or consumable makes it a supply, and a blank cell takes the file's kind. Any other word is asked about once, with Reagent, Supply or Skip these rows. A row marked sample waits for you, because samples go in their own file. You can change any row's kind in the review.

Locations that don't exist yet can be created during review, so a row can name a place your lab has not set up.

Each template has two example rows. The first fills in every column it can; the second fills in only what is needed now, to show what can stay blank. Both are examples to be deleted and replaced. They are not real data and not a suggestion of what any value should be. If an example row is still in the file when you upload it, the review marks it so you can replace it or leave it out.

You do not have to use the template. If your spreadsheet has its own column names, the matching stage handles that.

Stage 1: Match columns

Once your files are read, the first step shows How I read your file: which of your columns it thinks maps to which Labsistant field.

Your job here is to check it and correct anything wrong.

What you will see

  • Required columns. The minimum needed for a row to become an item. If one is unmatched, the step says which.
  • A confidence indicator on each match, showing how sure it is that a column matches that field.
  • Ignore this column on a column that will not be imported. Change its mapping if it should supply a field.
  • Two columns read as Lot number. Only one can be the lot number, so each offers to keep the other in the notes. The kept column's value is written to the lot's notes after its column name, so nothing in it is lost.
  • Storage location is marked as set at the locations step, because that is handled separately and more carefully. See Review and finish an import.

Change any mapping that is wrong. This is the cheapest place to fix a problem, because everything downstream inherits it.

A file that runs sideways

Some sheets list the field names down the first column and put one item in each column after it. When a file reads that way, this step asks: This file runs sideways, one item per column. Flip it? One click turns it to one item per row and matches the columns again. A column that only explains the fields, such as one headed What to enter, is left out, and the step names it. The file you uploaded is kept as it was; only the way it is read changes.

Review unit changes

The Units section shows each unit value from your file. Recognised conversions are pre-filled and marked Recommended, for example reading grams as g. Values needing a decision are marked Needs your input.

Read it rather than skipping it. It is telling you what it changed, and it is the place to catch a normalisation you did not want.

Units it does not recognise

If your file uses units Labsistant does not know, it says so and offers these choices per unit:

  • Change to one of ours, mapping it to a known unit.
  • Keep as typed, leaving your unit as written.
  • Skip, excluding the rows that use that unit from this import.

If you choose Keep as typed, also say whether the unit means Whole items or Measured amount.

Use this mapping stays greyed out until every unit it does not recognise is answered: changed to one of ours, kept with Whole items or Measured amount chosen, or skipped. Until then, that unit's card is marked Needs your input.

Whole items or a measured amount

Each reagent is treated as either whole items, such as boxes and packs, or a measured amount, such as a volume or a weight. Where this is implied by the unit you picked, the step says so and does not ask again.

Asking the AI for a closer read

If your column names are unusual and the automatic matching struggled, you can ask the AI to take a closer look at the file.

If it cannot be reached, the step tells you plainly that it could not reach the AI and lets you try again. It does not silently fall back to a guess.

Either way, the AI's reading is a proposal, not a decision. Everything it suggests lands in the same review you were going to check anyway, and nothing is created until you approve.

What is never guessed

Whatever route your data takes in, the import will not invent a compliance value. Lot numbers, CAS numbers, hazard levels, and expiry dates come from your files or from you, never from an assumption.

Rows missing something required are flagged rather than filled in. See Lot numbers, CAS, hazard and expiration.

Back to the Help Center

Try Demo Sandbox Start With Your Own Lab