Embla · Sync

Hello

Active file, suppliers and lines in the current sync set.

Lines and suppliers · all history

Monthly line and supplier counts from confirmed sync sets (line dates).

No dated lines yet. Confirm sync sets or load a dated CSV to build history.

Status

Sync set and last push

Current / last sync set

—

Confirm material suppliers on Import (step 4) to record a sync set.

Last Klappir export

—

Send to platform from Export to record a push.

Nothing asked yet. Answers appear here and stay for the session.

Answers come from a summary of the active file, not raw row-by-row lookups.

Share

One file, many customers

Share uses Klappir getActiveCustomers: customers who have requested data from this company. Upload the file. Kennitalas on that list route automatically. Everyone else waits until they request — no line review. Later Core can return a wider list on the same query.

Need a file to try? Download the sample (Acme 5508120020, Beta 5508120100, Gamma 5508120290, plus one unmatched).

This drop

Upload starts matching against customers who requested data from this company. There is no Review step.

Requested customers are planned for Core createFuel with their getActiveCustomers.id. Send stays 1→1 — this path never calls /api/transfer.

Not requested yet

Kennitalas in the file that are not on getActiveCustomers. They have not asked this company for data, so Share will not write them. Skip so they stay out of next month’s queue, or map only if the file id is an alias of someone who has requested.

Kennitala in file	Rows	Volume	Search list	Customer

Per recipient

Customer	Kennitala	Status	Core	Rows	Volume

Customers who requested data

This list is getActiveCustomers for the sharing company: customers who have asked you for data. A kennitala in the Olís file is not consent. Mapping a name is not a request. Later Core can drop that filter; Share will still match the file to whatever this query returns.

Vendor key not loaded yet. When SHARE_ROUTER_API_KEY is set, this list refreshes from Core.

Name	Kennitala	Core UUID	Status

Preview list (no vendor key)

Until the data-source key is set, you can preview matching with a local book. These rows are not Core consent.

Import

Import spend

Upload a spend CSV — classification runs automatically through the four mapping steps.

No file loaded yet. Review lines

1

UNSPSC mapping

Waiting for upload

2

NACE mapping

Waiting

3

GHG mapping

Waiting

4

Supplier material mapping

Waiting

Pipeline idle

0%

Upload a CSV to start the pipeline automatically.

Overwrite coarse source codes

On by default — catch-all codes (segment-level, or weak families such as transport / management / education) can be replaced if a more specific match is found. Turn off to keep every source code as-is.

Classify defaults are in Settings → Import.

GHG mapping (add-on)

Secondary to UNSPSC. Maps each supplier’s NACE code to GHG Protocol scope and Scope 3 category. Emissions are calculated on the Klappir platform after send — not here.

Suppliers to send

External suppliers start selected. Internal suppliers are flagged — include or exclude them before Confirm sync set.

	Supplier	NACE	GHG scope / category	Flags	Tx	Spend	Spend %

Coverage

Upload a CSV to see coverage.

Activity Show log

This file only

Clears classifications applied in this session. Codes that came with the file stay. Use this before re-running classify on the same upload.

Applied codes are provisional until review. Send to platform only includes suppliers left checked in step 4.

Review UNSPSC mapping

No file loaded

UNSPSC coverage

Material line items

The rows that can sync to Klappir. UNSPSC is shown as dashed code + name (e.g. 80-11-16-13 Temporary manual labor). Use Change to correct a code.

Status: Tagged by you — you set this code. Confirmed (source) — it came with the file. Auto-suggested (not applied) — a match exists but has not been applied. Needs mapping — no code yet.

A review flag appears when: there is no UNSPSC; the source code is not in the Klappir catalog; it is only a segment (too coarse for spend); or it is out of Purchased goods & services scope (travel, fuel, electricity). Source codes are never silently overwritten — we keep them and suggest an alternative. Unusual high/low activity (kg) and the highest/lowest spend lines are flagged for a look — they are not excluded from Send.

Suppliers in this sync set

Secondary view — NACE and spend share. Mapping happens in the line-item table above.

Loading classification codes from Klappir…

Classification coverage

Category breakdown

Classified spend (confirmed + tagged) broken down by family, then class, then commodity. Click a row to drill in.

Recurring terms

Words that recur across many descriptions, ranked by the spend behind them. Click one to filter the patterns below.

Description patterns

Descriptions from the same supplier are grouped by shared wording once numbers and details are stripped out. A pattern with several variants is the same underlying purchase described inconsistently, and where normalizing pays off most — but a supplier whose description never repeats still shows up here too, as a single-variant pattern, so every supplier can be tagged in one place, not just the ones with wording to clean up. Suppliers are listed once with all of their distinct patterns underneath, so a supplier with several different kinds of purchases doesn't get mistaken for a single one — check the pattern count badge before tagging. Tagging a pattern applies to every line in it, including ones that already have an existing code — use this to unify a pattern whose variants were classified inconsistently. Remove a tag any time to restore each line to its original state.

UNSPSC lookup

Search the full taxonomy (segment → family → class → commodity). Codes already used somewhere in this file are flagged with how often and for how much — a broad segment/family-level code used hundreds of times is a sign it's being used as a catch-all.

Search taxonomy

Line items

—

Total spend

—

Lines

Export

Sync history and Send live on the confirmed set below.

Download for review

Optional full line-item CSV. Keeps the original columns and adds mapping fields (source vs AI vs you).

Sync & transfer history

Confirmed sync sets with linked sends. Ready sets show Send to Klappir on the card.

No sync sets or transfers logged yet.

Send to Klappir

Push material suppliers only from this sync set via Goods & Services API. Same Unique IDs upsert safely on retry.

Sync set confirmed
Platform API configured
Legal Entity set
Material lines ready

Settings

Your account and Legal Entity. Classify and Send do not need anything else here.

Signed-in account and this browser. Legal Entity is already bound.

Account / workspace

Signed-in account

—

Workspace tenant: —

Send to Klappir

Legal Entity (UUID)

Managed by Klappir. You do not need to change this to Classify or Send.

This session

Start fresh

Clears the loaded file in this browser so you can upload again. Does not wipe learned labels or the Embla YTD / send history.

Reset history

Clears the Embla lines-and-suppliers YTD totals, confirmed sync sets, and platform send history for this workspace. Does not wipe learned labels or Legal Entity.

Classify

Classify defaults

These apply to every upload. Overwrite coarse codes is on by default on Import (catch-all source codes can be refined). Turn it off on that file if every source code must stay.

Family-level fallback

If no commodity or class fits, assign the best UNSPSC Family (level 2) instead of leaving the line blank.

Keep classifying until complete

Run another AI batch after each one until every remaining description is classified, or you press Stop. Turn off to run a single batch.

How Sync Studio reads, classifies, and learns

Upload starts a 4-step pipeline. Matches you (or the file) already know are reused before AI is called. After classify, external suppliers are selected; internal suppliers stay flagged off until you include them.

How we read the file A spend CSV is parsed for product or supplier text plus an amount. Asset, reference, Unique ID, date, unit, weight, and an existing classification code round-trip when present. Blank product/asset labels are filled from reference or supplier so classification and Send have enough text. Legal Entity UUID in a file is ignored — it is bound on this account. Missing Unique IDs are generated for export/send. Repeated header rows and total/summary rows are ignored automatically. The last uploaded file is kept in this browser so you can pick it up the next day; Start fresh or Discard upload removes it.
How we classify 1. UNSPSC — reuse learned labels for the same product text or product reference; remaining descriptions go to AI in spend-sorted batches. Family-level fallback (above) can assign a coarser code when a commodity is unclear. Firm source codes stay; coarse catch-all codes can be replaced when overwrite coarse is on (the Import default). Spend suggestions are family / class / commodity only — never segment. 2. NACE (add-on) — each supplier is mapped to an industry code. 3. GHG (add-on) — NACE maps to GHG Protocol scope / category. Emissions are calculated on the Klappir platform after send. High-CO₂ lines are flagged for review. 4. Suppliers to send — external suppliers start selected. Internal / intercompany suppliers are flagged off unless you include them, then Confirm sync set.
How we learn When a description is classified (from the file, AI, or a manual pick), the product text and product reference are saved to your Klappir account. The next file reuses those matches before calling AI. NACE by supplier is remembered the same way. A Nordic/Elins starter set loads automatically. This is the memory shown under Learned labels below — it survives new browsers; Start fresh does not wipe it.

Learning

Learned labels

Matched product texts are saved to your Klappir account so the next file reuses them before calling AI.

Preparing learned labels…

Account sync: waiting for sign-in…

File formats & column details

Minimum to import / classify Product Label or Supplier Label, plus a numeric amount (AMOUNT or a currency column such as dkk/eur/isk). Asset and reference are optional — blank product is filled from reference or supplier; blank asset is filled from supplier.
Minimum to send to Klappir A parseable date, amount or weight, product label (≥2 characters), asset label, and a confirmed UNSPSC. Legal Entity UUID lives on the account — do not put it in the file. Unique IDs are generated if missing. Rows can classify with less than this; Send skips them until the gaps are filled.
Export sheet includes Generated Unique ID (upsert key), supplier, date, weight, currency amounts, product, Classification Key, asset, and reference. No Legal Entity UUID. Emissions stay blank unless the staff opt-in is on — Klappir recalculates them after send.
Optional source columns Unique ID, date, unit, weight, and existing classification codes round-trip when present. A Legal Entity column is dropped on import.

Need access for a colleague? Contact Klappir.

Internal

Logic

Current mapping and send rules for Sync Studio. Subject-matter experts can propose a change; proposals do not go live until they are implemented in code.

Access needed

Hello

Lines and suppliers · all history

Status

Data path

One file, many customers

This drop

Not requested yet

Customers who requested data

Import spend

Suppliers to send

Coverage

Review UNSPSC mapping

Material line items

Suppliers in this sync set

Export

Download for review

Sync & transfer history

Send to Klappir

Settings

Account / workspace

Signed-in account

Send to Klappir

Legal Entity (UUID)

This session

Start fresh

Reset history

Classify

Classify defaults

How Sync Studio reads, classifies, and learns

Learning

Learned labels

APIs & connections

Platform API

Classify tuning

AI batch size

Libraries

Classification library

Learned labels (account)

Danger zone

Deep session caches

Clear sync & transfer history

Logic

Customers

Tenants

New customer

What to send the customer

Set-password links