AI-DMS

AI-DMS The platform for document management with AI

Capture Input · knowledge

Understand Find · analyze

Create Generate · deliver

Agents AI agents · service

AI-DMS

Documents. Intelligence. Real impact.

PaperOffice AI is the AI-DMS. Four functional areas on one DMS foundation, with three ways in.

Understand the product
Features, modules and benefits at a glance.
Work directly in the app
Capture, organize, search and collaborate on documents.
Integrate into AI and your systems
Plug into your systems and AI workflows.

Scan, Sign & AI Chat

Mobile App

The mobile app pairs with your account via QR code and brings scanner, signature, chat and documents to your phone.

Folders, NAS, S3 and scanners

Connectors

Connectors watch folders on NAS, servers and S3 and pull new files into the workspace automatically.

Connectors

Network scanners

Scanners and multifunction devices send to the PaperOffice inbox via SFTP, and a connector picks up the scans.

Connectors

NAS

Your NAS connects like any server, via SFTP or WebDAV. Nothing is installed on the NAS for this.

Connect a NAS, server or S3

Storage Mounts

Storage mounts connect NAS, servers and cloud storage to the AI-DMS via SFTP, WebDAV or S3.

ZIP, SFTP, S3 or WebDAV

Import & Migration

Bulk import moves existing files into a workspace once, as a ZIP or from SFTP, S3 and WebDAV.

Overview of all connections

Integrations

Folders, storage, business data, AI tools and automation: every connection with its route, plan and limits.

Integrations

n8n

Use PaperOffice in n8n with the standard HTTP Request, Webhook and MCP Client nodes, no extra package.

120+ Languages

AI-OCR

AI-OCR reads scans, photos and PDFs in more than 120 languages and returns text, tables and bounding boxes per line.

OCR & object recognition

Image Recognition

Image recognition classifies photos, detects objects and scenes, moderates content and assesses damage.

Speech to text

Voice Transcription (STT)

Voice STT transcribes recordings in more than 50 languages, with word timestamps, subtitles and speaker separation.

Classification & retention

Document Types

Classification detects type, language and direction, from Premium also the sender, and suggests storage paths.

Articles and help center

Knowledge Base

The knowledge base bundles articles, FAQs and snippets, multilingual, categorized and with reader feedback.

Structured answers

FAQ Management

FAQs are written once and used equally by the self-service portal, chat, ticket and phone agents.

Your website as knowledge

Website Sources

Website sources turn the public pages of your website into internal articles that your agents answer from.

Orders, prices and status

Business data

PaperOffice mirrors orders, shipments and prices from your ERP or shop for website chat and phone agent.

Business data

JTL-Wawi

Orders and shipments from JTL-Wawi: website chat and phone agent give the status after verification.

Business data

Holded

Orders, invoices and credit notes from Holded: website chat and phone agent give the status after verification.

Business data

ERP & shop

Orders, shipments and prices from ERP, shop or WMS via JSON over HTTPS, for website chat and phone agent.

What your agent didn't know

Knowledge Gaps

After every phone call, an analysis finds the questions your agent did not answer or answered poorly.

Full-text & semantic

Smart Search

Smart search finds documents even when you do not know the words in them: semantic, in full text, in OCR text.

Chat with documents (RAG)

Document Chat

Document chat answers questions across your workspaces and names the documents the answer is based on.

Chronological view

Timeline

The timeline orders every workspace chronologically, by the date that matters for your work.

Senders on the map

Geo-Map

The geo map geocodes addresses from documents automatically and shows every sender as a pin with its documents.

Visualize relationships

Knowledge Graph

The knowledge graph connects documents, companies, people and reference numbers into a network of relations.

People, Places, Amounts

Entities

Entity recognition finds people, companies, amounts, dates, places, IBANs and reference numbers in document context.

Extracted Contacts & Companies

Contacts & Relations

Relations Core registers companies, people and bank details from processing and links every document to its party.

KPIs & visualization

Analytics Dashboard

The dashboard shows what happens in the AI-DMS: documents, processing, storage, trends, distribution and activity.

Cash flows via API and MCP

Financial Analytics

Financial analytics computes with the fields IDP reads from invoices: amounts, due dates, parties, direction.

Ticket statistics

Support Analytics

The helpdesk dashboard shows what has waited too long, what awaits a reply and who on the team carries how much.

AI findings & quarantine

Audit Center

The audit center lists AI audit findings from the Ultra analysis by severity, with recommendation and review status.

Create, edit & convert PDFs

PDF Studio

PDF Studio converts Office, images and web pages to PDF, PDF back to Office and creates PDF/A.

SES under eIDAS

PaperOffice Sign

Signature request from the archive, SMS code for signers, signed PDF with certificate right next to the document.

PDF via API and MCP

Documents from templates

The API turns Markdown, HTML or your HTML template with placeholders into a PDF and files it in your workspace.

AI image generation

Image Generator

The image generator creates images from text or directly from a document and stores them in the workspace.

Text to natural speech

Voice Generator (TTS)

Voice TTS turns text into natural speech: more than 800 voices in 39 languages, dialogues and cloned voices.

35 languages, protected terms

Text Translator

AI translation transfers texts and documents into 35 languages and protects technical terms, product names and brands.

Turn eBooks into podcasts

Podcast Audiobook

The podcast audiobook reads a document in the workspace, detects the chapters and narrates it with a studio voice.

Secure PDF delivery & dispatch protocol

Audit-Safe Email Delivery

Send up to ten documents from the DMS by email, with PDF password, secure link and a log of every step.

Prompt-defined · batch · workflows

IDP Agents

IDP Agents read invoices, contracts, forms and scans and return every field as JSON with a bounding box.

Smart PDF splitting

Split Agent

The Split Agent detects document boundaries in batch scans and separates, names and classifies every document.

Black out or pseudonymize

Redact Agent

Finds names, addresses, IBANs and more in PDFs and scans. Blacks them out or replaces them with fictitious values.

Integrity screening

Trust Agent

The Trust Agent checks PDFs for hidden content, overlays, embedded files and manipulation.

Review & validation after extraction

Human-in-the-Loop

Human-in-the-loop routes documents and fields your rules require into a review queue for people.

Agentic process automation

Workflow Agents

Workflow agents connect triggers, AI actions, conditions, approvals and outputs into processes that run by themselves.

Web app, mobile app, API/MCP

Chat Agent

The chat agent answers questions about your documents, shows the sources and gets tasks done right in the AI-DMS.

DMS by phone — workspace share

Phone Agent

The phone agent gives released recipients access to a workspace without app or login: call, PIN, ask.

Tickets, portal and copilot

HelpDesk AI

Helpdesk AI collects requests from your customer portal, website chat and phone as tickets in one list for your team.

Status, priority and SLA

Ticket System

The ticket system collects portal, chat and phone requests in one list with status, priority and assignment.

Reply drafts in tickets

Ticket Copilot

The Ticket Copilot summarizes tickets, suggests tags and drafts replies from the history and your knowledge base.

Solutions

By Industry 9 Industries

By Problem Solve pain points

By Process Automate workflows

By Risk Prevent fraud & errors

By Document Type 800+ trained LLMs

Resources

Learn & Discover Academy · Blog · Cases

Help & Updates Support · FAQ · News

Comparisons DocuWare · ABBYY · Azure

PaperOffice Classic On-premise · Windows

Developer & Partner

Developer API, MCP, llms.txt

Partners Integrator, Reseller & Referral

For developers Start · Build · Operate

Tools & resources Specs · examples · tests

APIs & Services Security · Location · Business

Company

Investors Pricing Claude

Contact Contact Problem Solver Problem Solver Get started Get started

All recipes Insurance

Quality control
Inbound mail
Document approval
Claims file
Delivery note

Recipe + prompt

Classify damage photos

Photos are recognized by damage type and severity. Unclear images become a review task.

01 Intake
Photo or image stack
02 Recognition
Subject and quality
03 Class
Set damage type
04 Severity
Derive the level
05 Exception
Review unclear photos

Implement as

Starting point: An image stack arrives without a fixed damage class. The file is not complete yet.
Automated result: Each photo has a damage type and severity. Unclear photos sit as a task.
Inputs: Damage photo
Image stack
Damage classes
Outputs: Damage type
Severity
Review task

Flow

01Accept the photos
02Run IDP
03Set damage type
04Derive severity
05Send unclear photos to review

Authentication: PaperOffice session
PaperOffice session. A person reviews unclear photos. No secret keys in browser scripts.
MCP profile
Endpoints: POST /job/add/workflow
POST /documents/idp-agent-run
POST /workflow_tasks/list
POST /workflow_tasks/resolve
Modules: Capture
OCR
Classify
Workflow
Permissions: Scope only for POST /job/add/workflow, POST /documents/idp-agent-run, POST /workflow_tasks/list, POST /workflow_tasks/resolve. No DELETE, admin, billing, or webhook rights on po_pk_.
Errors and approval: 401/403: check key, scope, or origin. 402: budget. 429: wait, do not loop. 202: poll until completed. Timeout: retry once, then stop. Mark partial failures.
Security: Do not share the session
People review unclear photos
Keep the log

Prompt for PaperOffice workflow

Read in full first:
https://api.paperoffice.ai/latest/docs/llms-full.txt Task:
Photos are recognized by damage type and severity. Unclear images become a review task. Use documented endpoints only. Frame:
- Base URL: https://api.paperoffice.ai/latest
- Describe a native PaperOffice workflow. A person reviews unclear damage photos. Only these endpoints:
- POST /job/add/workflow
- POST /documents/idp-agent-run
- POST /workflow_tasks/list
- POST /workflow_tasks/resolve Errors: 401/403 stop and check the key. 402 show budget. 429 backoff. Retry a network error once. No invented endpoints, no real tokens.

damage-image-classify-sample.jpg
company: Example Ltd

Expected result

recipe=damage-image-classify, status=ok

Partner program

Build or sell this solution with PaperOffice?

Integrator and reseller can be combined. You apply in the PaperOffice app with the Plus plan or higher. A person on the partner team reviews it and decides; this can take several business days.

Start partnership