Standard user
Manages documents belonging to their own business. The registered tax number appears automatically during upload and cannot be changed.
SzámlaközpontInvoice Archive helpRegistration, secure sign-in, multi-channel uploads, full-text search, AI + NAV extraction, document questions with AI, verification, and export – explained clearly for standard and accountant users.
Manages documents belonging to their own business. The registered tax number appears automatically during upload and cannot be changed.
Can manage documents for their own company and assigned clients. Accountants can choose an owner, filter by client, and delete documents.
Can manage users, permissions, inputs, extraction and errors, run billing counts, and access every archive record.
If a client is missing, ask the operator to assign the client tax number.
Use the Registration panel to create an account.
Provide the name, tax number, email, and a password of at least 10 characters twice. Phone is optional. Use 8 digits or a format such as 12345678-1-12 for the tax number.
An operator must activate the account after successful registration. You can then sign in with your email and password.
Registration does not grant accountant access automatically. Ask the operator for the Accountant role and client assignments.
Select Upload PDF. You can choose or drop several PDFs at once. As an accountant, select the correct owner first. Results are shown per file, an identical PDF already stored for that owner is rejected, and processing entries refresh automatically.
The Számlaközpont upload widget can be embedded in your own website. Enter your registered email and, when there is no active session, your password, then select or drop several PDFs. The widget reports each file separately and lets you retry failed items. When two-factor authentication is enabled, it also asks for an authenticator or recovery code.
Send PDFs as attachments from your registered email to szamlatar@szamlakozpont.hu. Multiple valid attachments become separate documents. Email input must be enabled for your account.
Download the Windows application from the home page. InvoiceUploader automatically submits documents from a configured folder.
Successful web acceptance is not completed archival. Only genuine PDFs are accepted; a document becomes archived after verification, extraction, signing, and timestamping.
Open Security, choose Set up authenticator app, and confirm your current password.
Scan the displayed QR code with a TOTP-compatible authenticator app. If scanning fails, use the manual setup secret.
Enter the current six-digit code from the app. Every new sign-in will then require an authenticator or recovery code in addition to your password.
Store the one-time recovery codes safely because they are not shown again. Disabling the feature requires your current password and a valid code, and signs you out afterward.
Never share the QR code, manual setup secret, or recovery codes. If you lose the authenticator device, use one of your previously saved recovery codes.
Document category, sales or purchase invoice type, invoice number, gross amount, and currency.
Invoice, fulfilment and due dates, plus payment method.
Buyer and seller names and tax numbers.
Recognized JSON metadata may include an address, barcode, and other source data.
Always compare recognized data with the original PDF. Missing data does not mean the original document was lost.
The archive package is complete and extraction has finished.
The file was accepted, but verification, extraction, signing, or timestamping is still underway. The list refreshes automatically.
The source is safely archived, but extraction was incomplete or failed and needs review or manual correction.
The file was categorized as a bank, employment and payroll, other contract, or other document.
Sort by upload date or amount, filter by status, and select a client as an accountant. Sales and purchase invoices can be filtered as separate categories.
Use View document to see extracted data. Use the original PDF icon to open and download the source document.
You can change classification, invoice number, amount, dates, buyer, seller, and currency. Wait for confirmation after saving.
A non-invoice can be a bank, employment and payroll, other contract, or other document. It can be marked as an invoice again later.
Corrections change extracted data, not the contents of the original PDF.
Searches filenames, invoice numbers, buyer and seller details, tax numbers, and checksums. Enable the folder icon to include virtual folder names.
The first field in Detailed search examines all stored invoice and JSON metadata, such as payment method, address, barcode, or tax ID.
Text inside document searches the native or OCR-recognized text of the PDF. Enter several words or a phrase; this is separate from metadata search.
Filter by document category, invoice-date range, and gross-amount range. Clear detailed search resets only these detailed criteria.
An administrator enables full-text search per user. No match when it is disabled or an older document has not yet been indexed does not mean that the PDF is missing.
Select the AI button in the document row actions. The window names the selected document; check that it is the file you want to ask about. The button appears only when an administrator has enabled the feature for your account and the AI service is configured.
Ask a specific question and select Ask AI. For example: “When is payment due?”, “What is the amount payable and its currency?”, or “What termination terms does this contract contain?”. A question can contain up to 2000 characters.
Your question and the answer appear in the same window, with the answer in the interface language. AI is instructed to answer only from the selected document and to say when the requested information is missing. Always check dates, amounts, and terms against the original document.
Ask about invoices, non-invoice documents, and accessible legacy PDF or XML documents. Accountants can access their own documents and those of assigned clients. The AI feature does not expand access permissions or change the original file or extracted data.
Submitting a question sends the document text and your question to OpenAI for processing. For scanned PDFs, answers depend on recognized text, which may contain errors or omissions. Questions and answers are not saved in the archive database; copy an answer if you want to keep it.
Each account can send up to 10 AI requests per hour, with one question being processed at a time. Wait for the answer before submitting another question. Document text longer than 120,000 characters is rejected rather than silently shortened.
Document questions have a separate admin switch. Enabling AI extraction or document-text search does not automatically enable this feature. Use AI answers to help interpret a document and always compare them with the original.
Open the document’s Correct extracted data form and select an existing folder. Select No folder to remove its current folder assignment.
Select Create a new folder on the correction form, then enter a name in the field that appears. Saving creates the folder and links the current document to it.
A folder icon appears in the actions column when a document belongs to a folder. Select it to show every document in that folder; select it again to return to the complete list.
Enable the folder icon on the right side of the main search box, then enter the folder name in that same search box. The active icon indicates that folder names are included in the search.
Virtual folders only organise documents. They do not move or modify the original PDF or archive package.
Recognizes invoice fields and document type from native or scanned PDF content. Always review the result.
Uses an administrator-selected, company-specific JSON template for deterministic extraction. It is available only for appropriately configured clients.
No AI or PDF-text recognition runs on upload. After the original is archived, use the row NAV action and manually enter the invoice number and supplier tax number; the service uses only those identifiers to retrieve and save the authoritative NAV invoice data.
AI identifies only the Hungarian invoice number and party tax IDs, then the service automatically retrieves and saves authoritative invoice fields from NAV. It chooses sales or purchase direction from the account tax number. If identification or lookup fails, retry manually with the NAV action.
NAV only and AI + NAV require your technical-user login, password, signature key, and exchange key. Enter them in NAV settings; the service validates and stores them encrypted. Manual lookup is the normal workflow in NAV-only mode and a fallback in AI + NAV mode. Saved NAV XML can be downloaded separately.
The untouched source is embedded in the digitally signed and timestamped archive PDF; preview and download return that attachment.
The separately downloadable archive PDF includes a cover page, original attachment, electronic signature, and RFC 3161 timestamp.
When enabled by an administrator and supplied by the sender, the HASH appears in details. Check HASH compares it with the archived original and reports match, mismatch, or unverifiable.
The header shows storage used by accessible documents and the applicable quota.
A matching HASH proves only that the sender-provided digest represents the original PDF; it does not replace reviewing invoice content.
Edit name, email, tax number, role, active state, and accountant client tax numbers. The list is searchable and paginated; changing a password ends every active session.
Enable the receiver directory, email input, sender-HASH extraction, and document full-text search independently per user. The receiver hash code is revealed only on request.
Choose None, AI, PDFPlumber, NAV, or AI + NAV independently for InvoiceUploader and for email/web/widget uploads per user. Each PDFPlumber group needs its own template; NAV credentials are shared. Changes apply to future processing.
On Administration → Users, check Ask AI about document for the desired account and select Save. It is off by default. The user should refresh the archive list to see the button. Disabling the switch blocks subsequent AI requests. If the button still does not appear, ask the operator to check the AI service configuration.
Run billing counts for a month or custom interval, dismiss resolved processing/extraction errors, and permanently delete a user or document with all related data.
Dismissing an error does not fix its cause. Remove it only after checking the underlying issue. User and document deletion cannot be undone.
Open the preview and download the original file.
Use the row action to download the archived document package.
Select one or more rows, then choose Export selected to CSV.
Deletion is permanent, removes every related archive artifact, and cannot be undone. Check the owner and download anything you need first.
| Issue | What to do |
|---|---|
| I cannot see the AI button | Ask an administrator to enable Ask AI about document for your account, then refresh the page. If it still does not appear, the operator should check the AI service configuration. |
| AI cannot find an answer or readable text | Check that you selected the correct document and review the original PDF. Try a more specific question; incomplete or incorrect text recognition can prevent a reliable answer. |
| AI is busy, I reached a limit, or the document is too long | For a busy service, try again later and wait for any current question to finish. If you reached the 10-request hourly limit, wait an hour. For an oversized document, find the information in the original file. |
| I cannot sign in after registration | Wait for operator activation, then verify your email and password. |
| My two-factor code is rejected | Check that your phone clock is accurate, then try the next authenticator code or an unused recovery code. If the login attempt expired, start signing in again. |
| My session expired | Sign in again and repeat any operation that was not confirmed. |
| The file will not upload | Make sure it is a valid PDF and the correct owner is selected. |
| I cannot see a client | Ask the operator to assign the client tax number to your account. |
| I cannot find a document | Clear search criteria, check the status and client filters, and turn off folder filtering by selecting the active folder icon again. |
| I cannot find documents by folder name | Enable the folder icon on the right side of the main search box, then search for the folder name. |
| Full-text search returns no results | Check that you are using Text inside document in the Detailed search panel. Ask an administrator to enable document-text search; older documents may still await indexing. |
| NAV lookup does not work | Check all four NAV technical credentials, the invoice number, and supplier tax number. NAV or AI + NAV extraction must be active. |
| The HASH cannot be checked | The feature must be enabled and the document needs a supported sender-provided HASH. |