In your cloud, not in ours.
We install where your data lives: Azure, StackIT, IONOS or your own server.
OneMachine · Data-sovereign AI platform
OneMachine for data-sovereign AI in your infrastructure
The central AI machine for your employees: your own models, your own knowledge and your own workflows in your cloud, productive in about two weeks.
100+ models5 providers14-day demoEU AI ActGDPRMade in Germany
100+models available
5providers in parallel
~2 wksto production
100%data control stays with you
The platform
One interface for all your AI work.
Chat, agents, workflows, knowledge and governance in one place — securely in your infrastructure.
Only with OneMachine
From spreadsheet chaos to validated, queryable analytics.
Most business data lives in Excel: monthly reports, claims files, finance packs, sales exports, pricing sheets and operational workbooks. Yet these files are often hard to trust, hard to compare and difficult to query, because every team structures them differently.
OneMachine turns spreadsheets into governed analytical data. The platform ingests Excel files, detects the underlying tables, measures, dimensions, formulas and relationships, and converts them into queryable data assets. Instead of letting AI guess, OneMachine combines AI-assisted understanding with deterministic validation, evidence trails and human confirmation where needed.
- Tables, measures, formulas and relationships are detected automatically.
- Compare periods, validate formulas and detect changes across recurring files.
- Evidence-backed answers and reports — without rebuilding the logic every month.
- When the data is clear, OneMachine answers. When something is uncertain, it flags it for review instead of hiding the risk.
“OneMachine is built to answer when the evidence is strong, and to flag uncertainty when human review is needed.”
Less manual work, more trust.
Less manual preparation. Fewer spreadsheet errors. Faster reporting. More trust in every answer.
Platform
One machine for the entire value chain.
What you actually do with OneMachine.
Every model. Different per use case.
Five providers connected in parallel, over 100 models available. Plus your own models, which we host on your own server. You choose between security, cost and performance for each request.
Use case in action
This is how we do AI-powered sales ourselves.
Three minutes, one real run on OneMachine: before the first contact, the AI analyses where we can deliver value to a company with AI-powered decisions. When we reach out to you, this exact analysis is behind it.
- Qualification: the AI assesses which target companies are worth the outreach
- Automated lead research across CRM, web and LinkedIn
- Personalised emails and LinkedIn messages, ready to send
More than an AI tool.
Knowledge, workflows, connectors, transcription, data analysis, compliance. One machine instead of ten tools. Operational, not experimental.
Productive in about two weeks.
An onboarding document clarifies the requirements per cloud. Sign the DPA. We install in your infrastructure. You use it.
Sovereignty
Data sovereignty as architecture, not as marketing.
From document to answer: every step runs in an environment you control. This is what the data flow looks like.
Your data
OneMachine in your infrastructure
Models of your choice
Multi-tenant capable.
Roles, projects and teams can be controlled separately.
Your own models possible.
Open-source LLMs on your own server. Cost-effective and controlled.
Audit trail and evidence.
Every event logged. Compliance bundle exportable.
Token caps per user.
Spend limits for org and individual users. Live in the admin cockpit.
Connectors & compliance
Connects to your world.
SharePointOneDriveMicrosoft TeamsOutlookSlackCalendarERPCRMMCPREST APICustom connectorRAGVector databaseModel RouterGDPREU AI ActISO 27001SOC 2Audit trailSharePointOneDriveMicrosoft TeamsOutlookSlackCalendarERPCRMMCPREST APICustom connectorRAGVector databaseModel RouterGDPREU AI ActISO 27001SOC 2Audit trail
Stack
An example mid-market stack for AI.
OneMachine complements your existing Microsoft and data landscape as a controlled AI layer. The platform is installed in your infrastructure.
Users & Applications
Microsoft 365Teams, Outlook, Office
SharePointDocuments and knowledgeOneMachineChat, agents, projects
Identity & Access
Microsoft Entra IDSSO, MFA, Conditional Access
Roles & PermissionsTeams, projects, adminsOneMachine GovernanceModels, limits, logs
Orchestration & Automation
OneMachine WorkflowsVisual AI processes
AI AgentsTasks, routing, knowledgeOneMachine ConnectorsM365, Slack, ERP, CRM, DMS
LLM Models
Azure OpenAIEnterprise access
AnthropicClaude models
Mistral AIEU alternativeModel RouterUse case, cost, access
Data & Knowledge
PostgreSQL + pgvectorSelf-hosted knowledge
QdrantVector databaseOneMachine RAGSources, citations, retrieval
Storage & Documents
SharePoint, OneDrive, DMSExisting spaces
S3-compatible storageFiles, exports, evidenceEnterprise Knowledge LayerIngestion, versioning, embeddings
Logging, Audit & Monitoring
Cloud MonitoringAzure Monitor or native logsOneMachine Audit TrailPrompts, sources, models
Evidence BundlesGDPR, EU AI Act, ISO 27001 Sandbox & Rollout
Separate pilotTenant, project or VPC
Production operationTeams, roles, limitsInstallable in your infrastructure
Microsoft Azure
IONOS Cloud
Open Telekom Cloud
StackIT Cloud Your own serverRecommendation: The stack stays deliberately mid-market friendly: use your existing Microsoft environment, install OneMachine as the AI operating layer, and keep data and knowledge in your own infrastructure.
Packages
Three offers. One approach.
Start with a free demo or choose the right setup for your team.
Demo
For teams that want to test OneMachine in their own context before rollout.
Free14 days
- Guided start
- Demo access for evaluation
- A concrete use case as the basis
Team
Up to 100 users. For departments, project teams, production entry.
€699per month
- Up to 100 users
- Production operation for departments
- Admin cockpit and user management
Enterprise
For larger organisations, sensitive infrastructures and individual rollouts.
On requestindividually calculated
- Individual rollout planning
- Multi-entity and group operation
- Security and compliance alignment
All offers run in your infrastructure. Add custom connectors, support packages and individual SLAs as needed. Model token costs are passed through 1:1.
Onboarding document.
We clarify cloud, identity, data sources and first use cases.
Sign the DPA.
Data processing agreement under Art. 28 GDPR. With our documents. You sign.
Installation and go-live.
Installation, connection of the first data sources, go-live with your team.
Depth
Where others stop, OneMachine begins.
We are where you are.
Your data stays in your environment. No data migration to us.
Model flexibility.
Over 100 models across five providers, plus your own open-source models. Selectable per use case.
Per-user cost caps.
Token budgets per user and organisation. Live monitoring in the admin cockpit.
Source-faithful answers.
Every answer cites sources and shows the pipeline steps in a traceable way.
Visual workflows.
Drag-and-drop pipelines for repeatable team processes. Every step stays inspectable.
Privacy by design.
Audit trail, evidence bundle and human-in-the-loop are built in, not bolted on.
Per-user and org caps live in the admin cockpit.
FAQ
Frequently asked questions.
OneMachine is a data-sovereign AI platform that is installed in the customer's own cloud infrastructure. It bundles models, workflows, company knowledge, connectors, transcription, data analysis and compliance into one application, so companies run a central AI machine instead of many individual AI tools.
OneMachine is aimed at companies from around 50 employees that want to use AI productively and in a controlled way. The platform is especially relevant for FMCG, retail, industry, mid-market companies and groups with sensitive data, data-protection requirements or a need for an operational AI operating model.
Your data is processed exclusively in your own infrastructure, for example on Microsoft Azure, StackIT, IONOS Cloud or on-prem servers. Documents, embeddings, vector database, logs and model outputs stay in your tenant and in the cloud region of your choice.
Data-sovereign AI means that data is processed in an infrastructure controlled by the customer and does not sit with the provider. With OneMachine, documents, embeddings, logs, vector database and model access stay within the customer tenant; MindsMachines has no access to content data.
Yes. OneMachine is built on the principle of privacy by design and supports GDPR-compliant operation with data residency in the EU or Germany, a data processing agreement under GDPR Art. 28, role-based access, data minimisation, deletion functions and a complete audit trail.
Yes. OneMachine is designed so that operators can meet the requirements of the EU AI Act. This includes transparency, human-in-the-loop, risk classification per use case, documentation, audit trails and exportable evidence bundles for audits and internal governance.
OneMachine supports over 100 models across five providers, including OpenAI, Anthropic, Google, Mistral and Perplexity. In addition, your own open-source models such as Llama or Mistral can run on customer-owned servers and be selected per use case by security, cost and performance.
Yes. Open-source LLMs such as Llama or Mistral can run on your own server or in your cloud. This keeps sensitive workloads entirely in your infrastructure, while less sensitive use cases can optionally use cloud models via enterprise APIs.
RAG stands for Retrieval Augmented Generation and connects language models with your own knowledge base. OneMachine searches relevant document passages, evaluates sources, grounds the answer in context and shows citable evidence, so answers are traceable rather than made up.
OneMachine integrates with existing systems via standard connectors, Model Context Protocol and APIs. Supported services include SharePoint, OneDrive, Microsoft Teams, Outlook, Slack and Calendar; custom connectors for ERP, CRM and proprietary systems are possible.
Setup usually takes about two weeks. The onboarding document, DPA and technical requirements are agreed in advance; this is followed by installation in your infrastructure, connecting the first data sources, an admin training and go-live with your first use case.
The demo is free for 14 days. The Team offer costs €699 per month for up to 100 users; enterprise rollouts are calculated individually on request. Support packages, individual SLAs and additional services can be added as needed.
Last updated: 17 May 2026. Sources and references: GDPR Art. 28, EU AI Act, Model Context Protocol, Fraunhofer IESE on RAG.
See OneMachine in your world.
30 minutes. We show you the machine on your concrete use case.
Response within 24 hours · K67 Tech Hub Düsseldorf