Setup

First, your documents

  1. 1

    Connect your documents

    Connect the archives you already use — DocuWare, Google Drive, websites and URLs, or your custom line-of-business systems via webhook — or upload files by hand: PDFs, Word, Excel, text and images. Including scanned PDFs, electronic invoices and native Google files: Reperix extracts the text, with OCR when needed.

  2. 2

    Reperix organizes the information

    Your content becomes searchable by meaning, even when the wording of the question differs from the document. Permissions are yours to set with Reperix groups: you decide who sees what, and the same rule applies to every document, whatever source it came from.

All the sources you can connect →

The path of a question

What happens when you ask

Every question follows the same path. Here are the steps, in the order they happen.

  1. The question

    Write in natural language, as you would to a colleague.

    How much notice is needed to end the Rossi S.p.A. contract?

  2. Only the documents you can open

    Before it even searches, Reperix sets aside everything you could not open.

    • Contratto_Rossi_2024.pdf
    • Allegato_condizioni.pdf
    • Verbale_CdA_marzo.pdf not accessible
  3. The relevant passages

    From the documents that remain, only the passages about the question are taken — searching by meaning and by exact terms, and re-ranking them by relevance.

    • “…termination with 90 days’ notice…”

      Contratto_Rossi_2024.pdf, p. 4

    • “…renewal terms…”

      Allegato_condizioni.pdf, p. 1

  4. The answer, with its sources

    The answer is built only from those passages, and it shows where it came from, with citations that open the original document.

    90 days’ notice is required.

    • Contratto_Rossi_2024.pdf

The capabilities that matter

What makes the answer different

Beneath the simplicity of the chat, Reperix does precise work: answers grounded in documents, permissions respected, and search that understands what you mean.

  • Verifiable citations

    Every answer cites the documents it came from, with deep-links that open the original file in its source, at the right page.

  • You decide the permissions

    Reperix has its own group-based permission system: you assign who can see what, uniformly across documents from different sources. Each user, in answers too, only sees what you’ve granted them.

  • Searches by meaning and by words

    It combines semantic search with exact-term search and re-ranks results by relevance, so it finds the right phrase even when you asked it in different words.

  • Answers with charts, when useful

    If you ask for a chart, or when it helps to read the numbers, the answer includes one: built from the figures in the sources, never from invented values, always alongside the written answer and its citations.

  • Reads the hard documents too

    Scanned PDFs, electronic invoices, images: Reperix extracts the text, so even what you can’t search today becomes queryable.

  • One question, every source

    Archives, cloud, websites and custom systems answer together: you don’t have to remember which system the document ended up in.

  • On-premise or cloud

    Choose where the data lives: entirely inside your infrastructure, or as a managed service in a hybrid setup.

Beyond the chat

Share and integrate

The chat is only the start: you can bring the same answers to people outside the company and connect the systems you already use.

  • Shareable assistant

    A secret, expiring link that opens a chat over a subset of documents you choose: for customers, suppliers and collaborators, without giving them an account.

  • Webhook ingestion

    Your custom ERPs and document systems push documents to Reperix with a simple call, no dedicated development.

  • Websites and online docs

    Index pages and whole sites: public documentation or the company website become queryable, with citations to the live page.

  • Questions from your processes

    A workflow or a business app asks on your behalf, at the moment it matters, and uses the answer to decide the next step.

Shareable assistant · Webhook · Query API · All integrations

FAQ

The questions we hear most

Does Reperix replace my DocuWare or my document management system?

No. Reperix doesn’t replace your archives: it connects to the ones you already use (DocuWare, Google Drive, websites, or your line-of-business systems via webhook) and makes them queryable in natural language. Documents stay where they are, and citations reopen the file in its source system.

Do my documents leave my company?

In on-premise mode, never: documents and their analysis stay entirely inside your infrastructure. Even in a hybrid setup, when answer generation uses an external service it receives only the few snippets needed to answer, never the whole archive.

What file types and formats can it read?

PDFs (including scanned ones, via OCR), Word, Excel, text and images, electronic invoices, website pages and native Google files (Docs, Sheets, Slides). If a document contains text, Reperix makes it searchable.

How does it respect permissions?

You manage permissions in Reperix, with your own group system: you decide who can see which documents, and the same rule applies uniformly across every source (uploads, Google Drive, DocuWare, websites…). Before it even searches, Reperix sets aside everything the user couldn’t see, so nobody receives content they don’t have access to in an answer. Designed in from the start, not an add-on.

Where do the answers come from? Can I verify them?

Every answer is built only from the passages actually found in your documents, and cites them. The citations are deep-links: you open them and land directly on the original document in its source, at the right page.

Can Reperix create charts?

Yes. If you ask for a chart (say on a quarter’s trend) or when a chart makes numbers clearer, the answer includes one: bar, line or pie, built from the figures present in the documents, never from invented values. The written answer and its citations are always alongside, so you can verify the numbers. It is not a business-intelligence tool: it is the same document-grounded answer, also shown as a chart when it helps.

Can I give customers or suppliers access without an account?

Yes, with the shareable assistant: you generate a secret, expiring link that opens a chat over a subset of documents you choose. Whoever receives it asks questions without a Reperix account and only sees what you decided to share.

How long does it take to get started?

It depends on your sources and volumes. You can start right away by uploading a few files by hand, or connect an existing archive and let Reperix index it. In a demo we start from a real document from your company.

A note on data

Your documents stay yours

In on-premise mode, your documents and their analysis stay entirely inside your infrastructure. When answer generation uses an external service, it receives only the few snippets needed to answer, never the whole archive.

Learn more about security and deployment →

Try a question on your documents

In a demo we use a document from your company and show you the answer with its sources.