The quick copy-paste
A client email tossed into ChatGPT for a quick summary. Only after hitting send do you notice: the name, address, and case number were still in there.
Privacy Layer für sichere KI-Nutzung
MOM Shield erkennt personenbezogene Daten, Mandanteninformationen und Geschäftsgeheimnisse automatisch und ersetzt sie durch sichere Platzhalter — bevor sie an ChatGPT, Claude, Copilot oder andere KI-Tools gesendet werden. Die Originaldaten bleiben ausschließlich auf Ihrem Gerät.
The problem
A client email tossed into ChatGPT for a quick summary. Only after hitting send do you notice: the name, address, and case number were still in there.
A candidate profile as a PDF in Copilot — fast, because it was urgent. The compliance question comes afterward. The GDPR report catches it weeks later.
Your colleagues see how competitors move faster with AI. You hold back — because you don't have a workflow that reliably keeps sensitive data out.
Each of these situations shares one thing: a moment before the click where something should have been prevented. That's exactly where MOM Shield comes in.
Who MOM Shield is built for
MOM Shield is developed in the DACH region for industries where confidentiality is not optional. If you recognize yourself in at least one of these roles, the app was built for you.
Attorney–client privilege under § 43a BRAO. Briefs, contracts, and correspondence must not leak identifiable data to external AI services — not even for a quick wording suggestion.
Client data, balance sheets, tax files. AI assistance for explaining and structuring — without letting real names, IBANs, and tax IDs leak through.
Candidate profiles, personnel files, payroll data. AI helps with screening, evaluating, and writing replies — without CVs landing in plaintext at OpenAI.
Claim files, health data, contract numbers. Structure complex cases with AI without slipping into the special category under Art. 9 GDPR.
Customer documents, prices, terms, strategy papers. AI-assisted proposal drafting — without competitors ever getting near the originals.
Employees want to use AI, the DPO says no, leadership wants productivity. MOM Shield is the technical answer that satisfies all three.
Branchen-Fokus
Sehen Sie in unter 2 Minuten, wie MOM Shield in den Alltag von Anwälten und Versicherern passt — vom ersten Mandantengespräch bis zur Schadenmeldung.
§ 43a BRAO bleibt gewahrt. MOM Shield greift bevor irgendetwas das Haus verlässt — sei es ein Schriftsatz, eine E-Mail-Antwort oder ein Mandantengespräch.
Gesundheitsdaten, Schadenakten, BU-Leistungsprüfungen. KI hilft beim Strukturieren — und keine reale Personendaten landen bei OpenAI.
So funktioniert's
Sie ändern nichts an Ihrer Arbeitsweise — MOM Shield greift im Hintergrund. Vom Tippen bis zur fertigen KI-Antwort fühlt sich alles natürlich an.
01
Ganz wie gewohnt — E-Mail, Vertrag, Kandidatenprofil. Egal ob als Chat-Text oder Datei-Upload.
02
Personen, Adressen, IBANs, Aktenzeichen — automatisch erkannt und durch sprechende Platzhalter ersetzt. Optional bestätigen Sie die Regeln vor dem Senden.
03
Die KI antwortet mit den gleichen Platzhaltern — MOM Shield setzt auf Ihrem Gerät die echten Werte wieder ein. Lesbar für Sie, unbekannt für die KI.
Neu in v1.8 · Gamechanger
MOM Shield ist der erste Privacy-Layer, der nicht nur Text, sondern auch Bilder, Screenshots und gescannte PDFs automatisch maskiert. Integrierte OCR erkennt sensible Inhalte direkt im Bild — und ersetzt sie wahlweise durch sprechende Platzhalter oder schwarze Balken. Genau hier heben wir uns von allen anderen ab.
JPG, PNG und gescannte PDFs — Verträge, Ausweise, Rechnungen, Belege werden vor der KI-Verarbeitung lokal maskiert.
Sensible Inhalte werden durch strukturierte Platzhalter ersetzt — der Kontext bleibt erhalten, die KI kann weiterhin sinnvoll arbeiten.
Für maximale Privatsphäre: Inhalte werden mit schwarzen Balken vollständig unkenntlich gemacht — wie eine klassische Schwärzung, nur automatisch.
Two products under one brand
Both use the same masking engine, the same token standard, and the same security mechanisms. You can run them side by side — or just use whichever you need.
Extension for Chromium-based browsers. Masks instantly while you write in ChatGPT, Claude, Copilot, Gemini, and similar web AI tools.
Full Windows app for structured document work. Multi-user, audit log, license management.
Privacy & compliance
MOM Shield is not just a marketing promise — the architecture is built so that "data stays local" cannot work any other way technically.
MOM Shield is developed in Germany. No US-corporation data flows, no hidden cloud sync, no third-party telemetry.
Masking happens in your browser or your desktop app. Original data is never sent to a MOM Shield server — there simply isn't one.
Every masking, unmasking, and file access is recorded in the local DB with timestamp, user, and rule version. Exportable as CSV at the press of a button — for audits, DPOs, and management.
We deliberately speak of pseudonymization, not anonymization. The original values are recoverable — on your device, with your token. That matches the GDPR definition and is technically accurate.
Every commercial license is signed with an Ed25519 key pair. Tamper-resistant, verifiable, checkable offline. Suitable for ISMS-certified environments.
Admin and standard user are separated per installation. Self-registration can be disabled, a maximum user limit can be set. Password reset by the admin on your own server.
Frequently asked questions
No. There is no central MOM Shield server that could receive your data. Masking runs entirely on your device — in the browser or the desktop app. Updates and license checks happen offline.
Structured placeholders like [[PERSON_001]],
[[IBAN_001]], or [[ACCOUNT_001]]. The
AI understands the grammar and can work with the placeholders —
it just doesn't know who's behind them. When the response
contains the same placeholders, MOM Shield replaces them on your
device with the original values.
MOM Shield is a pseudonymization tool within the meaning of Art. 4(5) GDPR. Processing by the AI provider happens exclusively with pseudonymized data — that significantly reduces the risk compared to unencrypted transmission. A full legal assessment must still be made by your DPO in the context of your specific processing.
Up to 10 users per installation, each with their own account. One admin per installation. Larger setups (multiple locations, central policies) are on the roadmap — talk to us for a pilot.
The Chrome extension is free. The desktop app is currently in pilot phase — licenses are issued individually and are discounted or free during the pilot, depending on agreement. Regular pricing follows with general market launch.
The Chrome extension currently covers ChatGPT, Claude, Copilot (Microsoft), Gemini, and Grok. The desktop app is AI-agnostic — you mask locally and paste the masked text wherever you want. Local models (Ollama, LM Studio) work the same way.
Every license is signed with an Ed25519 key pair. Any tampering with the contained data (expiry, seats, plan) is immediately detected and rejected by the app. The public key is built into every app version; the private key stays exclusively with us.
For ISMS-certified or regulated environments (e.g., supervised insurers, hospitals, public authorities) talk to us about an enterprise setup: on-premises variant, custom rule set, signed builds. It's part of our roadmap and is being built out together with first enterprise customers.
Ready for safe AI use?
Try the Chrome extension for free. Or talk to us about the desktop app and an individual pilot setup.