Vikzaro Archive is a document archive you run on your own server, or that we host for you in the Vikzaro cloud. Scanned or uploaded documents are read with OCR, indexed into cabinets and found again with full-text search. When a document needs approval, the approver picks one of the decisions your company uses, and that decision becomes a status you can search for.
OCR, extraction and search run inside the application, and documents are not sent to outside AI services. The app is in English and Romanian.
Supplier invoices, stored by Ana Ionescu on 14 October
| Supplier | Papetăria Dealul Verde SRL |
| Invoice date | 10 Oct 2026 |
| Due date | 9 Nov 2026 |
| Department | Purchasing |
| Total | €476.00 |
Your decision on this step
An example document, with made-up data.
In just over 4 minutes the film follows invoices and contracts through the archive: how they come in and are read with OCR, how they are indexed and found again, and how approvals, redaction, versions and rights work.
Invoices, contracts and personnel files end up in shared folders, e-mail threads and binders. Vikzaro Archive gives each kind of document its own cabinet, with the fields you search by, the people who may see it and the rules for how long it is kept.
Upload them in the browser, drop them in a watched folder, send them to an e-mail inbox or photograph them with the phone. They wait in a tray while the text is read and index values are suggested.
You check the suggested values next to the document, click a word on the page to fill a field and store it in a cabinet. The suggestions learn from your corrections, per supplier.
Anyone with the right finds a document by its fields or by any word in it. Documents that need approval go to the right people, who see the document next to the decision.
Every cabinet is set up by your administrator. Five templates (supplier invoices, contracts, human resources, receipts, correspondence) give you a starting point.
Screens from the demo account, with made-up data.
An approver rarely has only two answers. In Vikzaro Archive the administrator sets the list of decisions for each flow, and each decision does something specific to the document.
When it is stored in a cabinet with a flow, or when someone presses “Send for approval”. Conditions on the fields choose the flow, for example invoices above a certain total.
It waits in their “To approve” list. They see the document next to the decision, read the notes and correct the fields the step allows.
Each decision has an effect: approve the step, reject, start again from the first step, send back to an earlier step or to the person who sent it. A comment can be required.
The status shows on the document, can be searched and can be written into an index field. A decision can also put a stamp with the name and date on the page.
| Decision | What happens | Status on the document |
|---|---|---|
| Approve | The step is done. After the last step the invoice is approved and gets the APPROVED stamp. | Approved |
| Reject to the department | The approval starts again from the first step, with a required comment. | Rejected to the department |
| Reject to the supplier | The invoice is rejected for good and gets the REJECTED stamp. | Rejected to the supplier |
| Clarification | It goes back to the person who sent it, with the question. When they answer and send it again, the approval continues at the same step. | In clarification |
You can rename these decisions, remove them or add others for each flow.
Run it on your own server, or let us host it in the Vikzaro cloud. On your own server, Vikzaro Archive is installed on a machine in your company or on a virtual machine you rent, and the documents, the database and the search index stay there. In the Vikzaro cloud we install it, look after it and update it for you. The application is the same in both cases.
Text recognition (OCRmyPDF and Tesseract), extracting the index values and the search engine (Meilisearch) run next to the archive, wherever it is hosted. No file and no text is sent to an outside AI service to be read, indexed or searched, and the cost does not grow with the number of pages.
Every file is stored on the archive's disk with its checksum, which is checked every week. A new version never overwrites the old one, and annotations and stamps sit over the page and leave the file as it is.
The Cloudflare Turnstile check on the sign-in page, Microsoft or Google sign-in if you turn it on, and the antivirus signature updates. The help chat inside the app, if it is on, sends an external AI model only the help text and the user's question, never data from the archive; you can turn it off. E-mail, webhooks, IMAP inboxes and external databases go only where you configure them.
Open-source OCR is weaker than commercial engines on poor scans and complex tables, and it does not read handwriting. Learned extraction needs a few documents from each supplier before it gets precise. OCR runs on the processor, a few seconds per page, so a large first import takes hours. Vikzaro Archive is not a certified archive with legal value; in Romania that needs an accredited provider.
Documents from the system you use now come across as files, together with their index in CSV or XML format.
From the system you use now, you export the files of the documents and their index, as a CSV or XML file, in a ZIP or a folder.
Every system exports a little differently, so we start with a sample of your export: we map its fields to the cabinet's fields and run a trial that stores nothing.
Each document is imported once, and an import that stopped continues where it was left. The files are processed like any new document, OCR included, so they can be searched by their text.
The index values and the files come across. Workflows, stamps and annotations from the old system do not: the approval flows are set up again in Vikzaro Archive.
In one of two places: on your own server, in your company or on a virtual machine you rent, or hosted by us in the Vikzaro cloud. In both cases OCR, extraction and search run inside the application, and documents are not sent to outside AI services. Tell us which option suits you when you ask for the demo.
A Linux server or virtual machine with Docker, for example 4 processor cores and 16 GB of memory, plus a data disk sized for your documents. Vikzaro Archive runs there as a set of containers: the application, the database, the OCR worker, the search engine and the antivirus. If we host it for you in the Vikzaro cloud, you do not need a server of your own.
On your own server, the archive works inside your network. Internet access is needed for the antivirus updates and for the services you choose to use: the Turnstile check on sign-in, Microsoft or Google sign-in and the help chat. On an installation reachable only from your network, the Turnstile check can be left off. In the Vikzaro cloud you reach it over the internet, from the browser.
On clear scans and PDFs created from software it reads the text well. On faded scans, photos taken at an angle and complex tables it makes more mistakes than commercial engines, and it does not read handwriting. You always see the suggested values next to the document before you store it.
Yes. When you correct a value, the archive remembers where it was on that supplier's documents, next to which label and in which part of the page. After a few documents from the same supplier, the suggestions come from what it learned.
Each cabinet has its own rights for people and teams: read, store, edit the index, delete, administer. A right can be limited to some documents, for example only those of one company. Redacted areas stay hidden from everyone without the right to see them, also in downloads and in the search.
No. It keeps documents with retention rules, checksums and an audit log, but it is not an archive with certified legal value. In Romania that requires an accredited provider.
Through the REST API, signed webhooks and autoindex jobs that read from your database (read only). There is no ready-made connector for a specific ERP; we look at what your software supports and set up the link with you.
Yes, if that system can export the files and an index as a CSV or XML file. We first test the import on a sample of your export, then import everything. See “Import from another archive system” above.
Please don't. The demo is for trying the application with made-up documents; its data are reset every night. To try it on your own documents, write to us and we'll set up a test on your server or in the Vikzaro cloud.
Leave your details and within minutes you get a demo account by e-mail, with an “Enter the demo” link. You get your own demo group, “<Company> (demo)”, where you are the administrator, and a second made-up login, aprobator@<company>.demo, to see the approver's side.
From 15 EUR/month per user Prices exclude VAT.
Already have a demo account? Enter the demo
Prefer email? Write to [email protected]
