Do we need to change our back office?

No. Syrve stays your POS and back office: we read its dictionaries and return documents into it. Team rights, warehouses and the catalogue do not change - we do not administer your system.

Does RestoAudit work without Syrve?

Partly - and we show the border honestly. Without a back office you still get intake from email, Telegram, PEPPOL, ZUGFeRD and XML, recognition of lines and prices, the purchase audit and reconciliation with the bookkeeper from their export. Menu cost from recipes, stocktake and automatic posting require Syrve.

Can we give read access only and not write access?

Yes, and that is the default state. We connect with a read-only user, parse the documents and show the exceptions. The right to post documents you enable separately, once you see the mapping is trained.

What happens to a new product that is not in the catalogue?

The line does not go into the back office and waits for your decision: link it to an existing product, or create the card in Syrve and refresh the catalogue. We do not invent items in your catalogue.

How is cost calculated if a recipe was changed mid-month?

We take the version of the recipe valid on the date of the document, not the latest one. Otherwise dishes cooked before the change would be costed with a recipe that did not exist on that day.

Do you post documents automatically?

Not by default: every invoice is approved by a person. Auto-posting can be enabled separately and only for a specific supplier - where at least 20 invoices have accumulated and nine out of ten went through without operator edits. Any exception on the document stops it.

What about exports to Jumis, 1C and DATEV?

Reconciliation with the bookkeeper runs from their export in xlsx, csv or PDF - their software does not need to be connected for that. The Jumis export is in progress, DATEV is planned, and we have no exchange with 1C. A developer can pull documents ready for posting through the partner API.

Where is the data processed?

In the European Union. Access to Syrve is a separate read-only user, write-back happens only on approval, and the processing terms are described in the DPA.

Product - Syrve integration

Syrve stays your POS and back office. We take away the typing

Integration is matching, not reading a PDF. «Chicken fillet chld.» at the supplier and «Chicken fillet chilled, kg» in your system are one product. A case of 12 is kilograms in the warehouse.

All integrations and API

Read-only by defaultWrite-back on approvalMapping is rememberedSyrve settings untouched

Two-way exchangelive

We read the dictionariesproducts, recipes, warehouses, salescontinuously

We match the linesthe supplier line with your product96% on their own

Disputed lines go to a persona new item, a doubtful unit4%

We post the invoicereceipt, return, credit noteon approval

Nothing is posted without approval. We read continuously, we write with your consent.

What the stock controller stops doing

Syrve counts what was entered into it. We close the place where it is entered by hand.

Before

Invoices are typed in by hand, line after line

The supplier line is searched for in the catalogue again

Cases are converted into kilograms on a calculator

A unit error travels into the cost of the menu

A return is entered as a receipt - it is faster that way

After

The document arrives parsed - all that is left is to approve it

The line is linked to your product once

Packs are converted to the unit of your warehouse

A doubtful line goes to a person, not into the back office

Return and credit note keep their own document type

96%

of lines pass on their own

after the first week of training the mapping

1 day

to connect

a read-only user and a connection check

changes in Syrve

settings, rights and the catalogue stay yours

41 h

of manual entry a month

no longer spent typing invoices

Demo data of a restaurant over one period. The exchange is the same, your numbers will be your own.

What we read and what we write - the full list

Dictionaries are read before the document is parsed, the invoice goes back after approval.

We read from Syrve

read-only

Products and units of measure

Recipes - the version valid on the date

Warehouses and their structure

Sales for the period

Before every document - so the mapping runs on the current catalogue, and the cost follows the recipe valid on that date.

We write to Syrve

on approval

The incoming invoice

Return and credit note - each with its own document type

A new supplier - on your approval

We never add a new item silently: the line waits for your decision

Every write is a human decision. Auto-posting is off by default and is enabled separately, one supplier at a time.

The decision is remembered for the supplier, not for the document

Linked once - the next invoice passes without a question. Until the supplier renames the item.

The link belongs to the supplier and the article

The training is not lost between invoices and does not depend on who took the delivery.

The unit is converted to yours

Case, box, multiplier - into the kilograms and litres of the warehouse. This is where money is lost most often.

The catalogue does not overgrow

A new item only with your decision. Otherwise in a month there are three «chicken fillets» with different units.

Invoice line - Syrve product

96% of lines pass on their own

mapping trained

At the supplierIn your SyrveUnit

At the supplier

Filejs vistas atdzisis

art. 22841

In your Syrve

Chicken fillet chilled

linked 14.08, remembered

Unitkg ← kg

At the supplier

Krējums 33% 1L x6

art. 10552

In your Syrve

Cream 33%

linked 02.08, remembered

Unitl ← box 6

At the supplier

Parmezāns 1kg

art. 30117

In your Syrve

Parmesan

linked 19.07, remembered

Unitkg ← kg

At the supplier

Alus 0.33 x24

art. 88204

In your Syrve

Bottled beer 0.33

linked 11.08, remembered

Unitpcs ← case 24

At the supplier

Tomati Datterino dzelt.

art. 41908

In your Syrve

new item

waiting for your decision

Unitkg ← ?

A supplier is created on approval, a new item is never added silently - the catalogue does not overgrow.

What works without Syrve and what needs it

An honest border: a back office is not needed for everything.

Works without a back office

Intake of documents from email, Telegram, PEPPOL, ZUGFeRD and XML

Recognition of lines, prices, units and VAT rates

Purchase price audit and supplier comparison

Reconciliation with the bookkeeper - from their export in xlsx, csv or PDF

Requires Syrve

Automatic posting of the invoice into the back office

Menu cost from recipes and sales

Stocktake reconciliation: book balance against the count

Month closing with the condition «write-offs reconciled with Syrve»

MCP

Ask your Syrve data in plain words

The same figures we pull from your accounting answer questions in Claude or ChatGPT. Not a file export - a conversation.

See how it works

A question instead of a report

What went up this month, why pasta margin dropped, how much was written off at the bar - answered from your Syrve data, not from the general knowledge of the model.

Read-only, and that is in the code

Twenty tools, all of them read. Changing an invoice, a price or a stock level through this access is impossible - no such tool exists.

Every call logged, access revocable

You see who asked what. Access is revoked with one button and expires on its own when unused.