Work with AI — without leaking confidential data.
mAIsk masks sensitive information before sending prompts to AI tools and restores it afterward — automatically and locally. Built for legal, healthcare, finance, education, IT, and anyone handling real-world confidential data.
One executable. Community by default. Pro unlocked with an activation key.
A simple safety layer between you and the AI.
mAIsk is designed for real work, not demos. It keeps sensitive data off the prompt while preserving a practical workflow for daily use.
Mask
Sensitive elements such as names, emails, URLs, IDs, IPs and other patterns are replaced locally with deterministic placeholders.
Use AI
Send the masked text to ChatGPT, Claude, Gemini, Codex or any tool you use. The AI never sees the original values.
Restore
Paste the answer back into mAIsk and restore the original data locally, in place, without cloud dependencies.
Built for professionals who handle real confidential data.
If your daily work includes names, emails, internal systems, identifiers, case numbers, or sensitive operational details, mAIsk makes AI usage more practical and safer.
Legal
Lawyers, conveyancers, legal assistants, compliance teams, contract reviewers.
Healthcare admin
Scheduling, referrals, back-office staff, patient-record workflows, medical administration.
Finance and accounting
Accountants, payroll, client file handling, invoice and reconciliation support.
Education
School administration, safeguarding-related notes, university support operations.
IT and operations
Engineers, sysadmins, support teams, log handling, ticketing and infrastructure work.
Consulting
Advisors who need AI speed but cannot expose customer, project or internal data.
Practical privacy, not theoretical promises.
The product is intentionally focused: stable masking, predictable placeholders, local restoration, and a workflow that makes sense under time pressure.
Deterministic placeholders
Stable, reversible placeholders using the format ⟦CATEGORY_###⟧ to preserve context and enable safe restoration.
Local-only processing
Masking and restoration happen on your machine. No cloud dependency is required for the core workflow.
Automatic + manual masking
Built-in detectors cover common data types. Manual masking handles the rest when judgment matters.
Dummy data substitution
Pro can replace sensitive values with realistic fake data when placeholders are too disruptive for the prompt.
Session-oriented workflow
Designed for actual iterative work with AI rather than one-off toy examples.
Activation-key unlock
One executable. Community by default. Pro unlocked through a local activation flow without turning the app into a SaaS.
Fast prompts, hidden risk.
- People paste names, emails, IDs and internal URLs under time pressure.
- Policy warnings are easy to ignore when work is urgent.
- The AI gets the real data.
- One bad prompt can create a serious confidentiality problem.
Same speed, safer inputs.
- Sensitive values are masked locally before they reach the AI.
- Users keep a practical copy/paste workflow.
- Restoration happens locally after the response.
- You reduce risk without banning AI usage altogether.
Simple enough to buy. Useful enough to keep using.
No subscription. No activation drama. Start with Community, upgrade to Pro when you need the full workflow.
For evaluation and light usage.
- Up to 10 placeholders per operation
- Basic masking and restore
- Built-in categories
- Local-only processing
- No dummy data substitution
- No custom categories
- No session persistence
For professionals working with real confidential data.
- Unlimited masking and restoration
- Dummy data substitution
- Session save and reload
- Custom categories
- Priority fixes and updates
- Single executable, activation-key unlock
Prompt leaks are not theoretical.
Public incidents and restrictions around generative AI keep showing the same pattern: people paste sensitive information into tools faster than policies can stop them.
A well-known example of employees pasting sensitive material into an AI system during real work.
ChatGPT history/billing exposure bugA reminder that even major platforms can have software bugs that expose information unexpectedly.
Why enterprises restricted AI toolsSome organizations limited generative AI use because staff were pasting code, strategy and confidential records.
Use AI more confidently with a safer workflow.
Start with Community or unlock Pro when you need unlimited masking, dummy data substitution and the full professional workflow.