Your clientsâ data doesnât belong in ChatGPT.
Now it doesnât have to.
Stript finds and replaces personal data in your documents (names, addresses, IBANs, everything) so you can use any AI tool without compliance risk. Runs entirely on your machine. Purpose-built for German and English documents.
Your data never touches a server. Not even ours.
See it in action
Stript detects personal data and replaces it with consistent pseudonyms, automatically.
This happens entirely on your machine. No cloud. No API. Just local AI.
The privacy gap in AI adoption
Itâs happening in your organization right now.
Right now, someone in your organization is pasting a client contract into ChatGPT. An HR report into Claude. A financial audit into Gemini. They know the risks but do it anyway, because the productivity gain is irresistible.
You canât ban AI. But you can make it safer.
How it works
From sensitive document to safe AI usage, entirely on your machine.
Upload or paste
Drop in a PDF, Word document, email, or scanned file, or paste text directly. Stript reads it locally. Nothing leaves your machine.
AI detects personal data
Stript reads your document the way a person would, locally. Names, emails, phone numbers, IBANs and more are found in context, not by blind keyword matching.
Review & confirm
Every detection is highlighted in context. Confirm, reject, or reclassify with one click. You stay in control.
Use any AI tool safely
Copy the anonymized text to ChatGPT, Claude, or any LLM. Paste the AI response back. Stript restores the original data instantly.
See it in action
From a sensitive document to safe AI use and back, four steps, entirely on your machine.
1. Add your document
Paste text or drop in a PDF, Word, email, or scan. Nothing is uploaded.
2. Review what's detected
Stript highlights names, addresses, IBANs, case numbers and more, and tells apart what is truly personal from public names like courts. You confirm each one.
3. Anonymize and use any AI
Consistent placeholders replace the data. Copy it into your LLM, or export a real .docx or .pdf that keeps the original layout. The mapping stays encrypted on your device.
4. Restore the originals
Paste the AI's answer back and your real data is returned, including a format-preserving .docx. Reversible by design.
Format-preserving export
Copy and plain text are always available. When you need the original back exactly, Stript can also export a real Word or PDF with the layout, tables, and formatting kept intact.
Real AI on your machine, not another cloud wrapper
Most anonymization tools cut corners. Stript doesnât.
Simple find & replace
Basic tools search for keywords and patterns. They miss names in context, catch false positives, and canât tell âBerlinâ the city from âBerlinâ the client.
Cloud AI service
Some tools send your documents to an API for analysis. That defeats the purpose: your sensitive data still leaves your machine and lands on someone elseâs server.
On-device NLP that understands context
Stript runs multiple AI models directly on your computer. It reads each sentence the way a person would, so it knows your client âLisa Weberâ is personal data, but the court âAmtsgericht MĂŒnchenâ isnât. No data leaves your machine, ever.
Same words, different meaning: Stript judges by the sentence.
Name
Date
Address
Three analysis layers. Zero cloud dependencies. Every model runs on your hardware.
Who is Stript for?
Built for professionals who handle sensitive data every day.
Data Protection Officers
Give your teams a compliant way to use AI, instead of fighting a losing battle against shadow IT.
Lawyers & Legal Teams
Summarize contracts, draft letters, analyze case law with AI, without exposing client data.
Consultants & Advisors
Process client reports and financial documents through AI tools without breaching confidentiality agreements.
Healthcare Professionals
Anonymize patient records, referral letters, and clinical reports before using AI for documentation or research, without compromising patient privacy.
IT & Compliance
Deploy a solution your teams will actually use, because it integrates into their existing workflow, not against it.
Built for privacy-conscious professionals
Everything you need to use AI without compromising sensitive data.
Your data never leaves your machine
Stript downloads AI models once, then runs fully offline. Your documents are analyzed by real NLP models running on your hardware, not sent to a cloud API. You can verify this: disconnect from the internet and Stript keeps working.
Understands context, not just keywords
Stript doesnât just match patterns. It reads language. It knows a birth date is personal data but an invoice date isnât, and that a clientâs name is sensitive while a public courtâs isnât. Three independent layers cross-check every detection: pattern matching, named-entity recognition, and contextual classification, all running locally.
Get your original data back in one click
Consistent pseudonyms like [PERSON_1] and [EMAIL_1] throughout your document. Paste the AI response back. Every name, address, and IBAN is restored exactly.
Works with your existing documents
PDF, Word, email (with attachments), scanned documents, or plain text. Upload a file or paste directly. Stript handles the extraction automatically, including on-device OCR for scans.
Bank-grade encryption for mappings
The link between original data and pseudonyms is encrypted with AES-256-GCM. Keys are stored in your operating systemâs secure keychain.
Catches what manual review misses
Benchmarked on German and English corpora of legal, financial and HR documents. Pattern matching finds structured data like emails and IBANs with 99%+ accuracy. AI catches the rest.
Detection quality you can measure
Benchmarked on real legal, financial, and HR documents in German and English.
>90%
Overall detection recall
>95%
Structured PII accuracy (email, IBAN, credit card)
100%
Local processing, no cloud
16+
PII types detected
How Stript compares
No more compromises between privacy and productivity.
| Feature | Stript | Cloud AI Services | Basic Tools | Manual Review |
|---|---|---|---|---|
| AI-powered detection | Partial | |||
| Context-aware classification | Partial | |||
| German language support | Varies | Limited | Manual | |
| Reversible anonymization | ||||
| Multi-format (PDF, DOCX, email, scans) | Text only | Manual | ||
| Runs 100% locally | ||||
| Encrypted mapping storage | N/A | |||
| >90% detection recall (benchmark) | Varies | Varies |
Frequently asked questions
Everything you need to know about how Stript handles your data.
Is my data really processed locally?
Yes, completely. Stript runs multiple AI models directly on your hardware, not through a cloud API. You can verify this yourself: disconnect from the internet after the initial model download, and Stript works exactly the same. No document content ever touches a server. The only network requests are the one-time model download, a lightweight update check on launch (plus the update download itself, only if you choose to install one), and (if you activate a license) a license check with our payment provider, as well as (only if you explicitly enable them in the settings) optional crash reports (technical error data); none of them ever contains document content.
How is Stript different from cloud-based or 'hosted in Germany' anonymization tools?
Cloud-based tools, even those hosted in Germany, still send your document to a server to process it. With Stript, the document content never leaves your device: detection and anonymization run entirely on your own hardware, and you can verify it in your network traffic. That is the core difference, on-device instead of hosted.
What data does Stript collect?
Stript collects no document content, no personal data, and no usage telemetry. Optional crash reports (off by default) send only technical error data with your explicit consent, never document content and nothing that identifies you. Encrypted anonymization mappings are stored only on your local machine.
How accurate is the PII detection?
Across our German and English benchmark corpora of legal, financial and HR documents, our three-stage pipeline achieves over 90% overall detection recall (measured 97â100% on the PII-sensitivity layer, depending on the corpus). Structured data like emails, IBANs, and phone numbers is detected with 99%+ accuracy. You always review and confirm before anonymization.
Which file formats are supported?
Stript supports PDF, Word (DOCX), email (.eml and Outlook .msg, including attachments), scanned or image-only PDFs (read on your device via OCR), and plain text. You can upload a file or paste text directly.
Can I use Stript in my law firm or company?
Absolutely. Stript is designed for professional use in regulated industries. The Enterprise plan includes self-hosted deployment via Docker, team management, and SSO/SAML integration. Your IT department stays in full control.
Is Stript GDPR-compliant?
Stript is built for GDPR-sensitive workflows. Document processing happens locally on your device. Your document content is not sent to Stript or to third-party AI providers by the app. Stript supports reversible pseudonymization and anonymization workflows. You remain responsible for reviewing the output and deciding whether it is anonymous enough for your use case. Website, purchase, licensing, and optional crash-report data are handled separately as described in our Privacy Policy.
What are the system requirements?
Stript runs on macOS (Apple Silicon, M1 or newer), Windows 10/11 (64-bit), and Linux (64-bit x86_64, as an AppImage or .deb package). Minimum 8 GB RAM, 16 GB or more recommended. You need between 3 and 20 GB of free disk space for the local AI models depending on your hardware tier, which download once on first launch and then run fully offline. Stript adapts to your hardware: on 8 GB it runs a lighter mode with smaller models, fully functional, but detection is slower and a little less accurate; with 16 GB or more it loads larger models and runs noticeably faster and more accurately. An optional GPU (Apple-Silicon GPU or NVIDIA CUDA) speeds detection up further. Either way, everything stays local on your machine.
How long does it take to process a document?
That depends on the document and your hardware. Larger documents and scanned PDFs take longer to analyze than short letters or emails, especially on machines with 8 GB of RAM. Stript shows its progress while it works, and you can review the first detections while the deeper analysis is still running.
Why is Stript a subscription?
Because the work doesnât stop at the download: the detection models keep improving, the app receives regular updates, and the license and update services run continuously. A one-time price couldnât fund that ongoing work. You can cancel anytime, and the free tier stays free.
How do refunds and cancellation work?
Purchases are processed by our payment provider acting as merchant of record. You can cancel a subscription at any time from the billing portal. It then runs until the end of the paid period. Details on withdrawal and refunds are set out in our Terms of Service. Read the Terms of Service â
How do I get support?
Email us at support@stript.io. We usually respond within two business days.
Simple, transparent pricing
Start free. Upgrade when you need more.
Free
Try Stript with no commitment.
- 5 documents per month
- All detection types
- PDF, DOCX, email, scanned documents & text
- 100% local processing
- Project management
- Full functionality: same detection quality as Pro
Professional
For professionals who handle sensitive data daily.
or âŹ470/year incl. VAT (save ~20%)
- Unlimited documents
- Word stays Word, PDF stays PDF, emails stay emails: anonymized exports keep layout, tables, letterhead and attachments
- AI answer back to a finished document: one click from LLM output to a restored file
- Scanned PDFs redacted directly in the page image, automatically verified
- Custom detection rules: terms and RegEx patterns, plus your own entity categories with their own placeholders
- Reusable detection profiles: save types, sensitivity, language, scan depth and active rules per matter type, switch in one click
- Compliance exports: mapping table (JSON/CSV) and a tamper-evident audit trail
- Email support
Enterprise
Self-hosted or cloud. For teams and organizations.
- Everything in Professional
- Self-hosted deployment (Docker)
- Team management
- SSO / SAML integration
- Custom entity types
- Dedicated support & SLA
- Coming late 2026
No credit card for the free tier. Cancel anytime. Prices include VAT, billed securely via LemonSqueezy. The free plan is the best way to try Stript before you buy.
Download Stript
Runs entirely on your machine: your documents never leave it.
macOS
Apple Silicon (M1 or later)
8 GB RAM minimum (16 GB recommended) · 3â20 GB free disk space for the local AI models
Download for macOSWindows
Windows 10/11, 64-bit (x64)
8 GB RAM minimum (16 GB recommended) · 3â20 GB free disk space for the local AI models
Download for WindowsLinux
64-bit (x86_64) · AppImage or .deb
8 GB RAM minimum (16 GB recommended) · 3â20 GB free disk space for the local AI models
Download AppImage or download the .debIf your browser blocks the download: open the downloads list, click the ââŠâ next to the file, then choose âKeepâ â âShow moreâ â âKeep anywayâ. Choosing âReport this app as safeâ there also helps us.
When you open the installer, if âWindows protected your PCâ appears, click âMore infoâ and then âRun anywayâ.
Stript is signed by Dr. Ingo Glaser, so youâll see the publisher name. These prompts are normal for newly released software and fade as more people download it.
After downloading, make the AppImage executable (right-click, Properties, Permissions, or chmod +x) and run it. No installation needed.
The optional libfuse2t64 package makes startup a little faster. A .deb package is available for Debian and Ubuntu.
On first launch, Stript downloads its AI models, roughly 3â11 GB depending on your hardware profile. After that, Stript runs fully offline.
Free: 5 documents per month, no account needed. Pro unlocks unlimited documents and format-preserving exports. See pricing
Stop choosing between productivity and privacy.
Download Stript and anonymize your first document today: free, no account needed.