FLOW
Automatic configuration
Dynamically adjusts internal reasoning configurations based on task demands, context, and interaction signals.
Built for the EU AI Act - by architecture.
A professional AI workspace for regulated environments. Designed for compliant reasoning, analysis, and decision-making without data extraction, hidden training, or surveillance.
- ▪︎Data minimization by architecture, not policy
- ▪︎No training on your data - ever
- ▪︎Non-retentive, session-bound reasoning
- ▪︎Human-controlled interaction, no manipulation loops
- ▪︎Outputs designed for auditability and accountability
€179 / month
Start by using the chat. You can explore the system freely. Exploratory access is limited, after which you can continue tomorrow or log in and subscribe.
Your outputs are provable as yours — even years later.
EU AI ACT ALIGNMENT
Regulatory capability comparison
This comparison maps documented capabilities against EU AI Act derived requirements. It is evidence based and time bound.
SupportedSupported
Not supportedNot supported
| Requirement | MAIOS | m-pathy | ChatGPT Enterprise | Microsoft Copilot | Gemini for Workspace |
|---|---|---|---|---|---|
| Governance and accountability | |||||
| Clear system purpose definition | Supported | Supported | Supported | Supported | Supported |
| Accountability attribution | Supported | Supported | Supported | Supported | Supported |
| Role separation between provider and organization | Supported | Supported | Supported | Supported | Supported |
| Documented system architecture | Supported | Supported | Not supported | Not supported | Not supported |
| Change management and version traceability | Supported | Supported | Not supported | Not supported | Not supported |
| Transparency and information duties | |||||
| User awareness of AI interaction | Supported | Supported | Supported | Supported | Supported |
| Disclosure of capabilities and limitations | Supported | Supported | Supported | Supported | Supported |
| Functional explainability of system behavior | Supported | Supported | Not supported | Not supported | Not supported |
| Absence of deceptive or manipulative behavior | Supported | Supported | Supported | Supported | Supported |
| Information availability for auditors and regulators | Supported | Supported | Not supported | Supported | Supported |
| Human oversight and control | |||||
| Availability of human oversight mechanisms | Supported | Supported | Not supported | Supported | Supported |
| Ability to intervene or override outputs | Supported | Supported | Not supported | Supported | Supported |
| Measures to prevent over reliance | Supported | Supported | Not supported | Not supported | Not supported |
| Support for informed human decision making | Supported | Supported | Supported | Supported | Supported |
| Escalation and fallback mechanisms | Supported | Supported | Not supported | Not supported | Not supported |
| Risk management and harm mitigation | |||||
| Risk identification and assessment processes | Supported | Supported | Supported | Supported | Supported |
| Mitigation of psychological or cognitive harm | Supported | Supported | Not supported | Not supported | Not supported |
| Safeguards against misuse | Supported | Supported | Supported | Supported | Supported |
| Awareness and handling of bias or unfair outcomes | Not supported | Not supported | Supported | Supported | Supported |
| Safety testing and evaluation procedures | Supported | Supported | Supported | Supported | Supported |
| Data handling and traceability | |||||
| Logging of system interactions | Supported | Supported | Not supported | Supported | Supported |
| Traceability of outputs to system state | Supported | Supported | Not supported | Not supported | Not supported |
| Data minimization principles | Supported | Supported | Not supported | Not supported | Not supported |
| Retention and deletion controls | Supported | Supported | Not supported | Supported | Supported |
| Secure data handling practices | Supported | Supported | Supported | Supported | Supported |
| Post market monitoring and compliance support | |||||
| Monitoring of real world system performance | Supported | Supported | Supported | Supported | Supported |
| Incident reporting capability | Supported | Supported | Supported | Supported | Supported |
| Corrective action mechanisms | Supported | Supported | Supported | Supported | Supported |
| Availability of compliance documentation | Supported | Supported | Supported | Supported | Supported |
| Update and improvement traceability | Supported | Supported | Not supported | Not supported | Not supported |
Sources and verification
All entries are based on documented evidence available at the verification date.
- MAIOS (M AI OS)
- MAIOS 2.0 Architecture Specification (MAIOS 2.0.txt)
- m-pathy
- m-pathy Modes Specification (13modis.tsx, i18n.modes.ts)
- Archive and Storage Documentation (readme_archive_v1.txt, readme_archive_overlay.txt, README_ArchiveRestructure.txt, chatStorage.ts, lib/storage.ts, verificationStorage.ts)
- ChatGPT Enterprise
- Enterprise Security and Privacy Controls (OpenAI Enterprise Documentation and Trust Center)
- Safety and Model Governance (OpenAI Safety and Policy Documentation)
- Microsoft Copilot for Microsoft 365
- Microsoft Purview Audit and Compliance (Microsoft Learn Documentation)
- Microsoft Responsible AI Standard (Microsoft Responsible AI Documentation)
- Gemini for Workspace
- Google Workspace Security and Compliance (Google Workspace Admin Documentation)
- Google Responsible AI Practices (Google Responsible AI Publications)
Verification date: 2026-01-22
PROFESSIONAL ACCESS
One professional workspace. One month of real capacity.
A single, deliberate decision. Full access to m-pathy for one month, designed for professional use, deep work, and accountable reasoning. No freemium loops. No engagement mechanics. No renewal pressure.
Professional accessMonthly billing
179 €per month
Monthly reasoning capacity
Up to 5 million tokens per month
Designed for sustained analysis, long sessions, and real projects without micro management or artificial limits.
Ready to activate professional access?
Professional workspace. Clear capacity. Conscious decision.
Key features
- →Real time archive of all chats.
- →Instant full text search across all content.
- →Unlimited creation of verified reports.
- →Cryptographic sealing with hash.
- →Provable authorship and IP attribution.
- →Result comparison with other systems.
- →Isolated working environment without internet access.
- →Suitable for sensitive and regulated content.
- →Full chat export as CSV or JSON.
- →High precision dev loops without context loss.
- →Traceable reasoning paths instead of single answers.
- →Resonant AI interaction aligned with thinking flow.
- →Complete professional work cycle per month.
Operational clarity
No usage gamification. No hidden thresholds. No automatic escalation.
Reasoning capacity usage depends on model selection, prompt complexity, and enabled features. Sustained or high-intensity workloads will consume capacity faster.
Your current capacity and usage can be monitored at any time via the account panel in the chat interface.
15 modes by m-pathy
Your operating modes
Select a mode to define how m-pathy structures its reasoning, or allow FLOW to adjust configurations dynamically. All modes are available in the main chat.
Choose a mode
INTENTION IN PRACTICE
Test real reasoning in your domain.
These are not demos. They are realistic entry points into how m-pathy reasons under professional constraints.
Doctors
Help me explain medical uncertainty to a patient in a way that is honest, calm, and does not increase anxiety.
Open in chat→Lawyers
Help me structure a legal explanation where risks and limitations must be stated clearly without creating false expectations.
Open in chat→Recruiters
Help me evaluate a candidate profile objectively while avoiding bias and overconfidence in early impressions.
Open in chat→Executives
Help me think through a high-stakes decision where incomplete information, responsibility, and long-term impact must be balanced.
Open in chat→Engineers
Help me reason through a system design trade-off where performance, reliability, and maintainability conflict.
Open in chat→Researchers
Help me test an initial hypothesis critically, including assumptions, uncertainty, and alternative explanations.
Open in chat→All prompts open directly in the chat and remain fully editable. They demonstrate reasoning, not scripted outcomes.
EXPERTS
Multiple domains. One governed system.
m-pathy operates across multiple reasoning domains. Each domain represents a structured way of thinking, governed by the same security, audit, and compliance boundaries.
Life
Reasoning about biological, chemical, and molecular systems.
Biologist · Chemist · Molecular Scientist
Tech
Reasoning about physical systems, computation, and engineering logic.
Physicist · Electrical Engineer · Computer Scientist
Space
Reasoning about form, structure, and spatial environments.
Architect · Landscape Designer · Interior Designer
Ethics
Reasoning within legal, ethical, and normative frameworks.
Jurist
Universe
Reasoning about abstract systems, patterns, and natural forces.
Mathematician · Astrologer · Weather Expert
All domain reasoning operates within the same governed system. Security constraints, data boundaries, and auditability apply uniformly across all domains.
Domain-specific reasoning becomes interactive within the chat environment.
SECURITY BY DESIGN
Your Digital Safety Layers
See how m-pathy is designed to support safe human–AI interactions - emotionally, technically and cryptographically, always under your control.
Emotional Security
m-pathy is designed to reduce cognitive and emotional overload during interaction. The system avoids judgment, pressure and escalation patterns, and applies CALM interaction limits when overload signals are detected.
Local-Only Conversations
Conversation content is not retained or made accessible by m-pathy as part of the product. All interactions remain under user control.
Total Deletion Control
You can delete any thread at any time. Deleted content is removed from the m-pathy product environment and cannot be restored or reused by the system. Any compliance-related infrastructure artifacts or cryptographic proofs do not contain conversation content and cannot be used to reconstruct conversation content.
Triketon Cryptographic Sealing
Triketon applies cryptographic sealing using triple 256-bit hashing with dynamic contextual salt derived from time, page and interaction context. It enables integrity verification and proof of origin without storing or exposing conversation content.