Architecture and migration
Sizing, target environment, data transfer, cutover and recovery are resolved before production operations.
WZ-IT plans, installs and operates Paperless AI as managed hosting, in your cloud or on premises. Depending on the target design, we also provide migration, secure network and identity integration, monitoring, backups, updates, integrations and further development.
Companies worldwide trust WZ-IT
The following are trademarks of their respective owners: Paperless-AI (the Paperless-AI project). WZ-IT is an independent service provider and has no business, partnership, or contractual relationship with these companies. We offer independent migration, installation, hosting, and operations services.
Paperless AI is an AI-powered extension for Paperless-NGX that enables automatic document classification, intelligent tagging, and semantic search using Retrieval-Augmented Generation (RAG).
With local or external model providers, Paperless AI can add classification suggestions and semantic queries. Data flows, model choice and business approvals are defined for the document corpus.
We install, host and operate Paperless AI for your business - either on our secure, privacy-focused infrastructure in Germany or other locations, or on-premise in your own environment.
We provide 24/7 monitoring, backups and maintenance for your Paperless AI instance. Human response times and support coverage follow the selected service level.
Analyses new documents and suggests titles, tags, document types and correspondents. Review and approval steps are used for sensitive collections.
Supports Ollama (Mistral, Llama, Phi-4, Gemma-2), OpenAI, DeepSeek, Azure OpenAI, Perplexity, Together.ai, LiteLLM, VLLM, Gemini and more.
Retrieval-augmented generation can provide suitable document passages for semantic queries. Quality depends on OCR, indexing, retrieval and the model.
Ask questions in natural language: 'When did I sign my rental agreement?' or 'Which documents mention my health insurance?'
AI-assisted suggestions for tags, correspondents and document types based on extracted content.
Search by meaning, not just keywords. Finds relevant documents even without exact matches.
Web interface for manual AI-assisted document analysis. Useful for sensitive documents requiring individual review.
Define rules for automatic processing. Disable prompts and apply tags automatically.
Self-hosted with Ollama integration - your documents stay on your server. Or use OpenAI-compatible APIs according to your privacy requirements.
Professional installation on your infrastructure - on-premise, cloud or hybrid
In your data center
AWS, Azure, Hetzner & more
Advanced architecture after technical and licence assessment
AI automatically analyzes and classifies documents with tags, types and correspondents
On-premise AI with Ollama - your sensitive documents never leave your infrastructure
Combine different AI models for optimal results with different document types
Find documents through natural questions instead of keywords: "When did I sign my rental agreement?"
Chat with your document archive and receive precise answers with source references
Perfect for companies with high document volumes and complex classification requirements
Secure access and access control for your installation
WireGuard, NetBird or Tailscale
Directly or through an upstream identity layer
Depends on application, edition and identity provider
Fail2Ban, Rate Limiting, IP Whitelisting
We set up secure VPN access to your installation - ideal for remote work and external employees.
Full-service installation with no hidden costs
Paperless-AI extends Paperless-ngx with LLM capabilities. We develop custom pipelines for your specific document types.
Via the Paperless API we control the entire workflow: Document ingest, tag assignment, metadata extraction.
Custom hooks for post-processing: After AI analysis, further actions can be triggered (archiving, notification).
We optimize LLM prompts for your document types: Invoices, contracts, correspondence - each type needs specific prompting.
How we implement Paperless-AI development in practice.
Incoming invoices must be manually transferred to accounting. Error-prone and time-consuming.
LLM extracts structured data (amount, VAT, supplier, IBAN) and pushes to DATEV/SAP.
Lawyers must check every new contract for critical clauses. Not feasible at volume.
AI analysis automatically marking risk clauses (liability, notice periods, jurisdiction) and flagging for review.
Incoming mail (scanned) must be manually assigned to the right department.
AI classifies letters by content and routes automatically: Complaints to support, applications to HR, invoices to finance.
We do more than provide an application. WZ-IT designs the technical architecture, integrates network and identity, operates the agreed scope and develops integrations when the standard product is not enough.
Sizing, target environment, data transfer, cutover and recovery are resolved before production operations.
SSO, secure access, internal systems and existing security components are integrated appropriately.
Updates, backups, technical monitoring and response paths follow a transparent operational scope.
APIs, automation and custom extensions can be delivered beyond basic deployment.
The exact scope depends on the application, edition, infrastructure and criticality. Vendor licences and non-standard components are quoted separately.
AI applications need controlled model access, knowledge sources, permissions and observability in addition to the user interface. We design these data flows as one coherent stack.
Teams, business applications and API clients use defined interfaces and endpoints.
TLS, firewall rules, reverse proxies or private network paths are designed around the platform's exposure.
Local accounts, SSO, directories, service accounts and emergency access are connected through clear roles.
Application logic, model access, roles and approved capabilities run in a controlled environment.
Local or external models, GPU/CPU resources, routing, limits and cost controls.
Documents, databases, vector search and traceable data paths for RAG and search.
Reviewed tools, APIs, logging, metrics and technical quality controls.
Models, extensions and data sources are not enabled indiscriminately. Permissions, data exposure, cost and professional oversight are assessed per use case.
A clearly defined operating scope instead of an opaque hosting flat fee.
We set up a test instance for you, usually on the next business day. No payment details required. After seven days it is deleted unless you continue.
Compute, applications, storage and response are shown separately. You can see what ongoing operations include and which requirements need a technical assessment.
We also design custom Paperless AI architectures, integrations and migrations. Contact us for a technical assessment.
One managed standard Paperless AI application is included in the Starter workload. Business and higher levels add a flexible operations allowance for planned work during regular service hours. Select compute, additional applications, storage and the appropriate service level.
A workload is one compute instance with the applications agreed for it.
One standard app per workload is already included. Additional dedicated servers count as separate workloads.
€7.99 excl. VAT per additional 100 GB per month, including daily encrypted offsite backup with 7-day retention. Billing is based on provisioned capacity, not actual usage.
We also implement custom backup schedules and retention periods. For example: daily backups retained for 30 days and one monthly backup retained for 12 months. Scope, storage requirements and additional costs are agreed separately.
+€299 excluding VAT per workload and month
Standard hosting runs on a single instance without high availability. Availability for custom architectures is agreed separately.
Zero downtime on request
Distributed operation with several instances, a database cluster and rolling updates, so that a node failure causes no interruption. Only for applications suited to it; design and price after technical assessment.
This page covers installation and the agreed platform operations for Paperless AI. Custom code, the access layer and special confidentiality requirements remain clearly separated responsibilities that can be added when needed.
Maintain the application
Keep custom extensions, interfaces and dependencies controlled, updated and supportable.
from €699.90 excl. VAT / month
View serviceSecure access
Add identity-based access and private network paths as a separate operational layer.
from €349.90 excl. VAT / month
View serviceOperate confidentially
For professional secrecy holders, define contractual, access and infrastructure requirements in a dedicated operating model.
from €499.90 excl. VAT / month
View serviceThree relevant adjacent paths instead of a long list of further products.
All prices are net and exclude statutory VAT. The offers are addressed to businesses.
View the overall systemEnquiry
Briefly describe the current state and objective for Paperless AI. We assess infrastructure, integration, and ongoing operations.
As an alternative to managed hosting in the data centre, WZ-IT provides the hardware, configures Paperless-AI, and handles hardening, monitoring, updates, backup and technical support. Access can be limited to the internal network or enabled through VPN and existing identities.
from EUR 349 excl. VAT / month · plus one-time provisioning and initial setup

Operations, models, data flows and integration
Paperless-AI is an open-source extension for Paperless-NGX. It adds AI-powered functions to document management: automatic document analysis, intelligent tagging, semantic search, and question/answer functionality based on your documents.
Including: automatic detection of new documents, AI analysis of content, automatic assignment of metadata (title, tags, document type, correspondent), semantic search, chat/Q&A function across documents, manual and automatic processing modes, and support for many AI backends (local or cloud).
Paperless-AI can connect OpenAI-compatible endpoints as well as local models, for example through Ollama. Which document content leaves the environment depends on the selected model, embedding and integration services.
Yes - Paperless-AI extends Paperless-NGX. Without a running Paperless-NGX installation, Paperless-AI will not work.
Typically via Docker or Docker-Compose: You clone the Paperless-AI repository, configure the URL and API token of your Paperless-NGX instance, select an AI backend, and start the container. Once the system is running, automatic analysis and chat functions are available.
Yes - both newly imported documents and existing documents can be analyzed by Paperless-AI and provided with metadata.
Yes - in addition to auto mode, there is a manual mode via a web interface where you can specifically select and review documents.
Significantly less time spent on tagging and classification, better consistency in metadata, faster and more precise search, easy retrievability, and a higher degree of automation - ideal for large document volumes.
Paperless-AI can be configured for a fully local setup, for example with Ollama. Whether documents and metadata leave the environment still depends on every selected model, embedding and integration service. We document these data flows and align operations with your privacy requirements.
Paperless-AI can run on your own infrastructure, for example alongside a local model server. To keep the data flow local, models, embeddings and other integrations must also be configured locally or under equivalent controls; legal compliance still depends on the concrete use case.
Yes - Paperless-AI provides a modern web user interface where you can configure settings, manually process documents, and use the AI chat.
Yes - Paperless-AI can also be installed on a Synology NAS (if Docker is supported) and thus works in many home or small business environments.
When using cloud models, document contents are sent to external providers at least temporarily - this can be privacy-relevant. Local models are more privacy-friendly but require sufficient hardware resources. Our AI Cube provides the necessary GPU power for local AI inference. A backup before the first AI analysis is also recommended.
For companies, organizations, or individuals with large document collections who want to automate their processes - e.g., archives, law firms, accounting, insurance companies, or anyone who regularly processes documents.
Yes - with semantic search and the chat/Q&A function, you can ask questions to the document archive ('When was our last invoice from company X?', 'Which contracts expire next year?' etc.) - not just simple keyword search.
Very flexible - you can set rules for which documents are automatically analyzed, define output tags, choose the AI model, and select manual or automatic processing.
Paperless-NGX itself supports PDFs, images, texts, and Office documents; Paperless-AI takes over their content and analyzes it - allowing intelligent management of various document types.
04.12.2025
Paperless-AI is an AI extension for Paperless-ngx that automatically analyzes incoming documents and assigns appropriate tags. Whether using OpenAI, local models with Ollama, or Azure...
04.08.2025
Paperless-AI is an AI-powered extension for Paperless-NGX that takes your document management to the next level. With automatic document classification, intelligent tagging, and semantic search,...
05.07.2025
Recent political developments show that it is becoming increasingly important to retain control over your own data. Paperless-ngx is an open source solution that allows...
These solutions are often used together with Paperless AI
These solutions offer similar functionalities and can be evaluated together
No risk: worst case, you leave with a clearer understanding of your project than before.


“WZ-IT's advice on our Azure migration was technically sound and completely non-binding right from the intro call - we took away a great deal.”
“WZ-IT moved our studio infrastructure from decentralised individual devices to a central platform: every site is securely connected via VPN, new devices are onboarded automatically and an entire site is provisioned from a template, without manual steps on location. What impressed me most is the breadth and depth of their knowledge: Timo and Robin are not a typical IT provider who sets up a server and leaves. The two of them think their way into highly complex infrastructure and software topics, work through every requirement we put in front of them, and build networking, provisioning and operations so that everything fits together in the end. WZ-IT is an excellent partner for complex software, network and architecture projects.”

Steve Kirchner
Managing Director, nextGYM GmbH

Whether a specific IT challenge or just an idea - we look forward to the exchange. In a brief conversation, we'll evaluate together if and how your project fits with WZ-IT.