Grow your business, not your back office.

Yes — the company memory is the brain of the platform. Every document is processed once, as soon as it arrives, and everything DocReveal learns from it is kept: a summary, context and key facts for each document, validated structured data, search by meaning and a knowledge graph of companies, people, amounts and their relationships across all documents. Chat, analytics, AI agents and integrations all draw on the same memory, so a document processed in January still answers questions in December — without being processed again.

Anything you would otherwise look up in the documents yourself — what a contract says about the notice period, which invoices do not match the agreed price, how much a supplier billed this quarter. You can ask about one document or across many at once. Every answer points to the exact place in the source, highlighted right in the document; it can include calculations, tables and charts, and you can see the steps that produced it. Personal data are hidden in answers, and an answer can be exported to Word, Excel or Markdown.

Digital colleagues that take over repetitive document work — sorting incoming invoices, watching contract deadlines or filing documents from a shared mailbox. You define them in the browser without code, much like briefing a new hire: what the job is, which mailbox, folder or system they may use, what they are allowed to do and how big their budget is. Ready-made templates help you start, and you test every agent on real cases before it goes live. Agents work over the company memory, use the same connections as integrations and keep a record of their work, so the same document is never processed twice. They come with the Agent workspace module.

You decide how independently each agent works: supervised, with safeguards or autonomously. Reading runs on its own, while actions that change something — sending an e-mail, labelling documents, writing to another system — can wait for a person’s approval, and the most sensitive ones always do. Approvals arrive in one inbox, every agent has a budget and run limits and can be stopped immediately, and each of its steps is recorded, so you can review it months later.

DocReveal treats extracted data as a proposal, not a fact. It says what it is unsure about, checks figures by adding them up where it can and validates every document against your business rules — for example that the line items add up to the total or that an invoice number never appears twice. Rules are put together by clicking, without code, and every version is kept. Mandatory fields and the confidence threshold are configurable, and documents that do not pass go to a person for review, together with the exact rule and values that failed.

Yes. Extracted data are ready for analysis straight away. Ask a question in plain words — for example whether any invoice arrived twice — and DocReveal turns it into a governed database query that you can review before it runs. Such queries only read data, never change them, and stay within your workspace. With the Management dashboards module, the whole team shares the same interactive dashboards with trends, financial indicators and process KPIs.

Yes. Your assistant connects to the company memory through the open MCP standard (Model Context Protocol) — you do not need the Agent workspace module for that. Each assistant gets its own key, limited to the capabilities it needs and with an expiry date, and you can revoke it at any time. Sensitive actions can wait for a person’s approval, and every call is logged and can be exported for audit. DocReveal charges per question; the model’s tokens are paid to your AI provider, which processes the content the assistant reads under your agreement with it.

Yes. Your data is stored in a secure cloud whose infrastructure is certified to ISO 27001 and SOC 2. Security is a first-class priority for DocReveal and we are continuously working on extending our certification portfolio.

Yes. Every company gets its own isolated workspace (tenant). Documents, extracted data, search indexes and the knowledge graph are separated from other customers on every layer of the infrastructure, so nobody outside your company can reach them.

No. AI models run through Amazon Bedrock within the EU; their providers do not receive your data and do not use them for training, and DocReveal does not train models on your documents either. The extracted data, summaries and knowledge graph that DocReveal builds from your documents stay in your own workspace.

Yes. Every processing step of a document is visible and can be re-run — the audit trail leads from upload all the way to the result. Business rules keep a record of who changed and published each version, every agent run and every call from an external AI assistant is logged, and integration runs can be reviewed and replayed.

When the cooperation ends, all data in your workspace are deleted completely and irreversibly — documents, extracted data, search indexes, the knowledge graph and connections to other systems. The deletion runs as an automated process that covers every place where your data are stored, not just the files.

Yes. DocReveal can process documents while keeping files in your own storage or cloud environment, including the cloud provider your company already uses.

Yes. DocReveal has its own AI module and can run on-premise, which also helps reduce future token-based operating costs.

With the Integration manager module, to a catalogue of more than 140 applications with over 1,500 ready-made actions — ERP and accounting systems, CRM, e-mail, spreadsheets, file storage, e-signature and project tools. Systems without a ready-made connector are reached through generic HTTP calls, and your IT team can describe an internal system as a connector of its own. Each service is connected once in Connection Hub; credentials stay encrypted, and the same connection serves both automation flows and AI agents.

No. Business users draw automation flows on a visual canvas. A flow starts with a trigger (a document was extracted, validated or needs review, an incoming webhook or a schedule) and continues with conditions and steps such as sending an e-mail, adding a row or creating a record. Live data from the document are inserted with a click. A flow can be tested before it goes live, every change is versioned, and the run overview shows each step with the option to replay it.

Whichever way suits you: upload them in the application one by one or in bulk, let the mail intake agent pick them up from a shared mailbox (Agent workspace module), or have another system send them through an integration flow (Integration manager module). An existing archive can be loaded in one go at the start.

Yes. A document is extracted once into one unified, versioned data structure, and from it DocReveal prepares outputs in several shapes at the same time — for your ERP or accounting system, reporting, automation flows and the company memory that chat and agents draw on.

Yes. We can quickly install a separate DocReveal version and configure it on any domain you choose.

Yes. DocReveal is ready to connect to Azure Entra or any provider that supports OAuth 2.0.

The core is always included: reading and understanding documents, the company memory, chat over documents, analytics, business rules and connecting your own AI assistants through MCP — with unlimited users and an interface in 24 languages. The add-on modules are the Agent workspace (your own AI agents), the Integration manager (automatic transfer of data to other systems) and Management dashboards (interactive dashboards for management).

No. The calculator is indicative and non-binding — it shows the prices of the individual items. The exact price is set in a quote based on the Process analysis. Prices exclude VAT.

By pages: one page is one page of an uploaded document; longer documents are counted page by page. Each page is paid once — further reading and questions over a processed document are not charged again.

By actual usage, billed monthly with a detailed report. Loading a page into the company memory is paid once. Questions from your own AI assistant through MCP are charged per question, and the model’s tokens are paid to your AI provider; chat with the DocReveal AI model is priced on request. Agent runs are billed by agent hours, depending on how demanding the task is.

No. The core includes an unlimited number of users. Only the Management dashboards module is licensed per user.

The price list is set in Czech crowns. EUR and USD are an indicative conversion at the ČNB rate stated next to the calculator; the currency and amount of the contract are set in the quote.

Basic operations maintenance and technical support come with every delivery: availability on business days from 9:00 to 17:00, e-mail and a notification portal for requests, monitoring, backups, updates and a monthly overview of usage and incidents. Rule changes, new document types, mapping changes or small integrations are covered by the optional Basic, Active or Full support packages, with reserved capacity and response times from three business days down to four business hours. The application also has a built-in help center, and video guides are available on this website.

Yes. The same Terms of Service apply to companies and to consumers. Consumers also have the statutory rights described in Section 11 of the Terms – in particular the right to withdraw within 14 days and rights from defective performance.

All legal documents are linked in the footer of this website: Terms of Service, Privacy Policy (including the Cookie Policy), Data Processing Terms (DPA), Security Overview and information under the Data Act. The documents are currently available in English.

The user or customer must have the right to upload documents and have them processed. If documents contain personal data of third parties, the customer is usually the controller and DocReveal acts as processor under the DPA.

Documents are stored and processed on Amazon Web Services in the region you choose: Europe (Stockholm, Sweden) or the United States (N. Virginia). AI models run through Amazon Bedrock within the same geography, so EU data stays in the EU and US data in the US; the model providers do not receive your data or use it for training. Another AWS region is available on request within 24 hours. Details are in the Security Overview and in the list of sub-processors linked from the DPA.

No. AI outputs are assistive and may contain inaccuracies. Users should review them before using them for accounting, legal, tax, HR, business or other important decisions.

Payments are processed by Stripe. DocReveal never receives or stores full card numbers; it keeps only the subscription, invoice and payment-status data needed to manage your plan. Invoices and payment history are available in the Stripe customer portal, which you open from Billing Manager in the application.

Data is stored in a secure cloud certified to ISO 27001 and SOC 2. Details about the infrastructure, subprocessors and controls are described in the Security Overview and DPA, and the certification portfolio is being extended over time.