AI for your documents. On your servers.
MadeForPrivacy is the platform your AI agents work on: with documents, mailboxes, databases and your systems, on your infrastructure or in your own cloud. Three data zones define which AI services may see which data. Every run is listed in the tracker with its cost, down to the cent.
30 minutes · remote · no obligation
from €399 per month incl. support · €10 per user
- Runs on your premises or in your cloud
- Data zones per agent
- Cost per run, always visible
- Developed in Ludwigsburg, runs on your server
Built for mid-sized companies
In use at companies that have no AI department and don’t need one.
MadeForPrivacy is already in use at mid-sized companies in industry and IT, with a dedicated contact person, predictable costs and data zones that fit how they work.
In brief
See MadeForPrivacy in action.
Video transcript
AI in your company? Absolutely. But your data should stay in-house.
Hi, I'm Neva, your personal assistant. You give me a task, I plan it step by step and assign the right agents to it.
Take an order, for example. Whether it arrives as an email, a screenshot or a PDF, order processing extracts it and creates an order card with the fields your business needs.
You review it, correct it where needed and hand it over to your ERP with one click.
The data zones decide which AI service may see what: Local only, GDPR-compliant or Unprotected. And every step is logged, including its cost.
MadeForPrivacy runs on your server, with the AI service of your choice. From 399 euros per month. See it for yourself in a live demo.
Runs on the infrastructure you already have
Why MadeForPrivacy
AI in business rarely fails because of the AI.
It fails because of data protection, costs and integration. That is exactly what MadeForPrivacy was built for: a platform that runs agents on your infrastructure instead of sending your data to someone else.
- Data in third-party clouds.Customer data flows to external AI services. Data protection becomes a matter of contracts rather than technology.
- Flying blind on costs.Every prompt costs money, and nobody sees what for. At the end of the month you get a bill with no breakdown.
- Siloed tools.A chatbot alongside your systems: data gets copied back and forth, processes stay manual.
- Data protection as architecture.Data zones define which data may be processed and what leaves your company. The platform enforces this technically, not just by contract.
- Full cost control.Every run with steps, tokens and cost down to the cent, per agent, per run and over time.
- Integrated, not bolted on.Agents run as tool pipelines connected directly to APIs, LDAP/AD, SQL and Microsoft 365, right where your processes are.
Business value
Value comes from the process, not from the chat window.
AI has arrived in companies; for most of them, the returns have not. Large surveys show why: what's missing is rarely the model, but integration into workflows and a clear calculation of what a task costs with AI and what it saves.
of CEOs have seen neither lower costs nor higher revenue from AI so far.
PwC Global CEO Survey 2026 · 4,454 CEOsof companies do not yet have the capabilities to achieve tangible value from AI at scale.
BCG 2024 · 1,000 executivesof the obstacles lie with people and processes, only around 10% with the algorithms.
BCG 2024 · respondents' assessmentSelf-reported survey data. These figures describe the market and are not MadeForPrivacy metrics.
A process, seen through to the end
Incoming invoices, incoming mail, orders: the agent processes the mailbox or file share you connect, and its tools write the result via API, SQL, email or Microsoft 365 to wherever the process continues. Nothing ends up in a chat window, and nobody retypes anything.
Affordable to run
A low-maintenance Docker stack on your infrastructure, with predictable licensing per server and user. The AI is only called for the steps that need it, and simple steps run on small models. Token costs stay low and are billed directly by the provider.
Every task has a price tag
The tracker records every run with steps, tokens and cost down to the cent, per task, per agent and over time. So you know what each process costs and can compare it with what the work costs you today.
Data protection
Here, data protection is architecture, not a statement of intent.
Every piece of information belongs to one of three data zones. Every agent step runs in exactly one of them, and MadeForPrivacy enforces the rules technically before a single byte leaves your company.
Local only
Examples: internal data, employee data, contracts.
Never leaves the company. Processing happens exclusively on your own infrastructure, including the AI.
GDPR-compliant
Examples: customer data, order data, correspondence.
Processed only by services approved for it. Each step sees only the data it needs, and every step is logged. The result in this zone is GDPR compliance.
Unprotected
Examples: public content for marketing, social media, creative work.
May use any AI service. This zone deliberately holds nothing that needs protection.
- No data leaves without a zone rule
- Every agent step is logged
- On-premises or in your own cloud
- Data protection by design, not retrofitted
From local models to GPT-6 on Azure (EU) to Claude and Gemini: which AI providers fit which zone. You sign the contract with the provider yourself.
The platform
A chat on the outside. An operating system behind it.
MadeForPrivacy works like a chat, but it is more than a chatbot. Behind it, agents run as tool pipelines with data zones, tracking and cost control.
Like a chat, with an operating system behind it
Start agents, follow their runs and get results right in the conversation. Neva, the personal assistant, delegates to specialists.
Every run is traceable
Steps, data flows, tokens and cost per run, live in the tracker, for continuous optimization.
Data flows with zone logic
Local only, GDPR-compliant or Unprotected: you decide which data may go where, and MadeForPrivacy enforces it technically.
Cost per agent, run and token
Token-efficient thanks to tool and process logic: the AI is only called where it adds value. You can track what it costs over time.
Connected directly to your systems
API, LDAP/Active Directory, SQL, Microsoft 365, email: no data is copied by hand, and processes run where they belong.
Your agents
No off-the-shelf agents. Each one fits your business.
Ready-made agents are the starting point, adapted to your processes, systems and data zones. And when you need something new, you build it yourself: in the client, without code.
Ready-made agents, tailored to you
Invoices, emails, documents, support, translations: for typical tasks there are agents that work out of the box. Together, we adapt prompts, tools and data zones to your workflows, not the other way around.
Create your own agents right in the client
Every agent is a tool pipeline: clearly defined steps, structured inputs and outputs, optional validation. You create new agents in the client, choose the data zone and permissions and go live when you're ready.
Adjust existing agents at any time
Refine a prompt, swap a tool, change the data zone, give a group access: changes take effect immediately. The tracker shows every run with steps, tokens and cost. That is how you improve agents while they are in production.
By SMEs, for SMEs
Manageable effort. Full control in operation.
A predictable rollout, low-maintenance operation with Docker and a dedicated contact person, from the first pipeline to everyday optimization.
Data zones decide which data may go where
%
of runs traceable: steps, tokens, cost
dedicated contact person from rollout to operation
Manageable costs
Predictable effort for the rollout, straightforward operation afterwards. Efficient use of AI and cost tracking keep running costs under control, too.
Simple operation
Easy to maintain thanks to Docker: services are cleanly separated and can be deployed and updated reliably. Less complexity, stable operation.
Personal support
A dedicated contact person for rollout and further development. If needed, we also support you with operations and integrations.
Frequently asked questions
What companies ask us first.
Short answers here. Everything else we can discuss in a call.
Does MadeForPrivacy run entirely on our own premises?
Yes. MadeForPrivacy is Docker-based and runs on your own infrastructure or in your own cloud. You decide per data zone which AI services an agent may use, from local models via Ollama to approved cloud services such as Azure OpenAI. With local models, the platform can also run entirely without an internet connection if you want; tools that need the network, such as web search, are then not available.
Which AI models can we use?
Local models via Ollama, GPT-6 and other models via Microsoft Azure (EU), plus OpenAI, Anthropic (Claude), Google Gemini and OpenAI-compatible endpoints. You approve each service for a data zone and select it per pipeline step. That way, an agent for internal data runs entirely locally, while another one uses a large cloud model for public content. You sign the contract with the provider yourself; the full list is under Integration.
What happens to our data during AI processing?
Every piece of information belongs to one of three data zones. Local only: never leaves the company. GDPR-compliant: processed only by services approved for it; each step sees only the data it needs and is logged. Unprotected: deliberately public content. The platform enforces the rules technically, before a single byte leaves your company.
How do we keep an eye on costs?
Every run is logged with steps, tokens and cost down to the cent, per agent, per run and over time. Because tool and process logic control the flow and the AI is only called where it adds value, consumption stays predictable. You can find the license models under Pricing.
How quickly can our first agent be up and running?
For typical tasks such as invoice or email processing, document analysis or customer support, there are ready-made agents that we adapt to your processes together: data zone, tools, prompts. After that, you create new agents yourself in the client or change existing ones whenever a workflow changes.
How does MadeForPrivacy connect our systems?
Directly via API, LDAP/Active Directory, SQL databases, Microsoft 365 and mailboxes. No data is copied back and forth by hand, agents run where your processes are. Learn more under Integration.
Live demo
See MadeForPrivacy live.
In a short call, we talk about your use case and show you MadeForPrivacy live, including possible next steps.
Short call · remote · no obligation
Philipp & Philipp · Founders